Today, the same code only runs in Virginia. Marina's slowest loads are about 1.5× faster — and across the app, a typical load is now about half a second.

Edge compute is supposed to be the fast option. For us it was the slow one.

A one-line config change, a Hong Kong bug that started it all, and an API that ended up twice as fast as a side effect.

Let's dig in 👇

A bit of background

Two quick definitions, since they'll matter for everything that follows.

A Cloudflare Worker is, roughly, a tiny piece of JavaScript that runs inside Cloudflare's network. Rather than spinning up a server in one data centre and pointing DNS at it, you hand Cloudflare your code, and they run it in hundreds of locations around the world. You get global scale and (usually) low latency essentially for free, with none of the usual ops overhead. It's a really lovely abstraction.

A PoP — point of presence — is one of those locations. Basically a Cloudflare data centre, with hundreds dotted across every continent, mostly in major cities. When a user makes a request, it lands at the PoP nearest to them, and that's where your Worker runs.

So the default behaviour is: every request runs at the PoP closest to the user. Which is exactly what you want if your Worker isn't making many external requests — if it can do its work locally and ship a response.

It is not what you want if your Worker spends most of its time waiting on a database halfway across the world.

The setup

The Shelf app is powered by our API, which is implemented as a Cloudflare Worker. The Worker itself is mostly glue: most of what it does is reach out across the network. It talks to Supabase (a hosted Postgres), writes to Kinesis, reads from S3, and fans out to a handful of third-party APIs.

A lot of those reads and writes go to things hosted in AWS us-east-1, in Virginia — our Postgres, Kinesis, and S3, in particular.

In plain English: when Marina opens the Shelf app, a Worker somewhere in the world wakes up, makes a handful of round-trips to Virginia, and ships her back a response.

If that "somewhere in the world" happens to be São Paulo, each round-trip to Virginia takes about 100 milliseconds. A typical request makes five or six of them. That's already half a second of pure network time before we've even started executing the database queries themselves.

Going looking for the problem

This didn't actually start as a latency investigation. It started in Hong Kong.

A couple of our Shelf Intelligence features quietly stopped working there. They use AI models that aren't available in Hong Kong, and because the model calls were originating from whichever Cloudflare PoP was closest to the user — for Hong Kong users, that's a PoP in Hong Kong — those requests got blocked at the source.

When we went looking for a fix, the obvious one fell out almost immediately: run the API from a known region instead of from wherever the user happens to be. That solves the Hong Kong problem on its own — the model call now originates from us-east-1, where it isn't blocked.

But once I was already in the traces, a second thing jumped out. Pulling an hour of data from Grafana Tempo:

• Our Worker had run across 65+ PoPs on 6 continents in a single hour.

• Only ~3.6% of invocations ran in PoPs near us-east-1 (EWR, IAD, CLT). The other ~96% were executing far from the data they needed.

• Tail latency mapped almost perfectly to distance from Virginia.

So picture Marina, opening her feed. A Worker in GRU (São Paulo) wakes up, sequentially round-trips to a database in Virginia five or six times, then shoves a response back across the Atlantic. Of course it was slow.

The Hong Kong fix and the latency story were the same fix. Pin the Worker to us-east-1, and Hong Kong gets its features back and Marina probably gets a faster feed. How much faster, I genuinely didn't know — I had a guess of maybe ~50% from eyeballing the traces, but it was a side effect I was curious about, not the thing I was solving for.

Why distance hurts so much

Here's the shape of the cost.

Every API request from the Shelf app makes five or six sequential round-trips to the database before it can respond. When Marina's Worker is in São Paulo and the database is in Virginia, each of those round-trips eats ~100ms of pure network latency. That's ~500ms per request, sitting there on the wire, doing nothing. Multiply by all the database calls a typical screen needs and the picture sharpens fast.

When you move the Worker physically close to the database, each round-trip drops from ~100ms to well under a millisecond. Five round-trips of <1ms is a rounding error. Five round-trips of 100ms each is half a second.

Call this the data-gravity tax 💸 — every byte your code wants from your database costs you a round trip times the distance, and edge compute does nothing to help you pay it. The closer your code lives to your data, the smaller the tax.

The fix

Cloudflare gives you two ways to override the default placement:

1. Smart Placement profiles your Worker for ~15 minutes after each redeploy and infers where to run it. It's the right call when you have multiple backends in unknown locations.

2. Explicit region pinning lets you just say "always run here." It's the right call when you know exactly where your backend lives.

We know exactly where our backend lives. So we picked option 2.

"placement": { "region": "aws:us-east-1" }

That's the whole change. One line of JSON.

The results

• p50: 1.5s → 0.5s (−65%)

• p95: 8.6s → 4.3s (−50%)

(Across 9 hours of pre-cutover traffic and 24 hours after.)

I had been expecting maybe a 50% improvement on a good day. The p50 fell by a full two-thirds. And it wasn't a one-day fluke: the improvement has held ever since.

The app now loads and feels meaningfully faster, especially outside North America — exactly the cohort that was hurting most.

Who actually benefited

The improvement wasn't uniform. Pinning Workers to us-east-1 was a bet that Worker → database round-trips dominated total latency — and the bet paid off most for users furthest from Virginia (p95):

• 🇹🇭 Thailand: 6.1s → 1.9s (3.2× faster)

• 🇧🇷 Brazil: 5.3s → 3.6s (1.5× faster)

• 🇲🇽 Mexico: 2.4s → 2.4s (flat)

The further from Virginia you were, the bigger the win.

Oh — and Hong Kong got its Shelf Intelligence features back. The original problem.

What we didn't expect

Users had been quietly giving up.

Cloudflare logged about 6,600 cancelled requests in the 9 hours before the deploy — meaning the client closed the connection before our Worker could respond — and zero in the 24 hours after.

The iOS side told us what those cancellations actually were: users opening the Shelf app, waiting for it to load, getting frustrated, and closing it before anything appeared on screen. Before the fix, roughly 460 users a day were doing this. After: about 260 users a day.

So the fix didn't just make things faster for the users who waited. It recovered around 200 users a day who had been silently bouncing out of the app before it could load — users we'd never have known about as anything more than a cancellation log line.

When this doesn't apply

One important caveat before you go off and pin all your Workers.

This trick only works because our backend is single-homed in a known region. If your reads are CDN-cacheable, run at the edge — that's literally what the edge is for. If your backend is genuinely multi-region, look at Cloudflare's Smart Placement instead. We're trading "fast for everyone in theory" for "fast for everyone in practice" because we know exactly where our data lives.

The data-gravity tax doesn't go away — but you can choose where to pay it.

Takeaway

The edge is a good default, but not always the best solution. Cloudflare's "run near the user" default is brilliant for most workloads — which is exactly why it's invisible. Nobody questions it, nobody re-examines it once they've deployed, nobody profiles it. But "good for most workloads" doesn't mean "good for yours". Pay attention to the data-gravity tax your platform is quietly charging you. Go look at the defaults you accepted years ago.

That's it for now — go pin those Workers if you have the right kind of workload, and have a great week! 👋

Join the livestream!

We’ll be chatting about a few topics:

• Personal AI needs to know you and act for you. That requires a rich, current understanding of you and permissioned access to your services.

• Apps are becoming content. What’s the future of infinite interfaces.

• Your context should travel between apps and agents, under your control. What does building the shared layer for memory, connectors, service access, and permissions entail?

  Join the livestream!

He is right. But the sharpest version of this paradox is not playing out inside the enterprise tenant…

A person pays for intelligence twice as well. Once with money, and another with their personal information. To make an assistant useful you must feed it your questions, your plans, your drafts, your doubts. The better you want it to perform, the more of yourself you have to reveal.

Models learn from exhaust. A firm’s exhaust is prompts, tool traces, and corrections, distilled into institutional know-how. A person’s exhaust is more intimate. It is the messages you’ve sent, the questions you’ve asked, the album you replayed all week, the book you abandoned on page 40, the correction that says no, that is not me. Every correction is distilled into self-knowledge, the kind no advertiser could ever buy.

Over time the asymmetry skews further. The system learns what you love, what you avoid, how you decide, and what you will pay for. You learn almost nothing about what it has learned. Somewhere on someone else’s infrastructure sits the most complete portrait of you ever assembled, and you cannot read it, correct it, or take it with you.

Nadella writes that in consuming intelligence you are creating intelligence, and that what you create should belong to you. Nowhere is that truer than for a person. This is your particular intelligence in Hayek’s sense: the knowledge of time, place, and circumstance that no one else can hold. We have an older word for it. Taste.

Taste is a lifetime of experience compressed into a signal about what you will love next, and memory is the basis of identity. Whoever holds your memory holds a working draft of you.

The irony he flags lands harder here too. These models were trained on the public output of millions of people under fair use. Then the terms turn around: the provider reserves the right to learn from your private usage while granting you no equivalent right to the learning. When learning flows in only one direction, economic value converges toward the owners of the learning infrastructure rather than the people generating the signal.

We have already run this experiment once. In the social era, platforms modeled your taste in order to sell your attention. Your data created the value, and you got a feed. The AI era raises the stakes, because this time it is not your attention being modeled. It is you.

As Alex Karp put it: “What the technical customers want is control over their compute, their models, their data stack, and their alpha. They want to know they own the means of production, and it’s not being transferred to someone else.” Swap “technical customers” for “people” and the sentence holds. People just do not have a procurement department to negotiate it.

So the trust boundary Nadella prescribes for the enterprise has a personal equivalent, and we need it more.

One place where your memory, taste, corrections, and context accumulate and compound together. Legible to you. Editable by you. And a hard boundary across which nothing crosses, not even the exhaust, without consent. This is not privacy as abstinence. The point is not that the internet should know less about you. The point is that it should know you on your terms, and do things for you because of it.

That requires unbundling the stack: the agent you talk to, the model underneath, and the layer that holds who you are should not be one company’s black box. The model provides the intelligence. You should decide what it knows. And, importantly, none of this can feel like homework. It must be fun. Firms buy trust boundaries through procurement; people adopt what is fun. Whatever hands the learning loop back to individuals has to be expressive, delightful, and accessible to the average human, not just the tech-savvy among us, so that control becomes the byproduct of something you would use anyway rather than the pitch.

In the platform era, companies accumulated your data. In the AI era, they accumulate the learning of you. Protection has to evolve accordingly, from guarding information to guarding the mechanism through which you are learned. There are a few things every person should be able to expect:

• Control: Your taste and memory is your private eval. It defines what you like and what you’re like, what “good” looks like for you, and no general benchmark can. Retain ownership of your memory, the record of you that you’ve nurtured, along with the right to grant or revoke access to any slice of it.

• Capability: Keep a living model of yourself inside your own boundary, structured, legible, and correctable, something like a genome of the self that you author. Models should learn against your real life without absorbing it.

• Choice: Decouple being known from any single assistant. Ask yourself: if the model you talk to every day were taken away tomorrow, would you still be known? The generalist model is rented. The veteran that knows you must remain yours.

• Cost: Portable context lets you bring yourself to whichever model is best and most efficient for the task at hand, without sacrificing quality. There should be no loyalty tax on being understood.

• Compound: Bring these together and you have what I think will change the internet: a continuous learning loop that belongs to you. Taste compounds over a lifetime. The understanding you invest in an agent should compound the value of your life, not the valuation of someone else’s model.

In other words, a person should be able to use intelligence without giving up themShelves. let’s see who spots that.

Personal Context Infrastructure is the most important evolution in personal technology since social media.

The social internet gave us a way to create online - to participate through our content. Social networking first used this power to bring connection, mapping our friends, relationships, and communities. Then social media grew to eclipse it, turning friendship into followership and connection into performance - fueling the creator economy, influencer commerce, and entire cultural movements.

Personal Context Infrastructure will bring the personal internet—the intelligent internet. It provides the missing layer that transforms software from generic tools into personalized intelligence that understands you, adapts to you, and acts on your behalf.

Today’s AI looks magical in demos but feels shallow in practice because it works with fragments of your story. You’re forced to keep prompting and retraining - repeating yourself across disconnected services. Context is gathered in pieces, yet the whole of who you are - and what you need - remains just out of reach.

The Current Fragmentation Problem

ChatGPT doesn’t know your communication style, your calendar app doesn’t understand your energy patterns, your streaming service doesn’t know what you’re working on this week.

Each service starts from zero, forcing endless onboarding processes that never quite capture who you really are. This fragmentation fundamentally limits what AI can do for you - the most powerful AI capabilities require rich context for understanding preferences, predicting needs, automating decisions, and taking autonomous actions.

The problem is that today the stitching together of all this context falls on us. We type, search, compare, and repeat. That is why online shopping feels like work instead of a red carpet experience. It’s 2025: software shouldn’t feel like filling forms.

Personal Context Infrastructure is the intelligent layer that aggregates and interprets the full picture of who you are across your digital life - under your control. Software should serve your life first, not a company’s quarterly goals.

Context is what software knows about you right now; memory is what it carries forward and refines over time. Personal Context Infrastructure connects the two, allowing software to remember you - not just respond.

It turns scattered data fragments into a portable digital identity that flows seamlessly between services on your terms. Think of it as your personal context server - a living memory that belongs to you, continuously enriched by your experiences and guided by your preferences - your LEARNINGS.md for digital life.

The Three Fronts of Change

Rethinking the Business Model

Today’s platforms build moats by hoarding your context—locking it inside walled gardens optimized for time on platform and ad yield. In the process, they fracture your memory across services. Personal Memory Patforms demand new economics: platforms that earn a share of your identity, not your time, and business models that thrive on user-permissioned context liquidity instead of lock-in. See our 2024 paper on the topic here.

Rebuilding the Technical Architecture

Personal Context Infrastructure must be user-controlled by design. Your memory lives on your turf, accessible only on your terms. The context can go where you choose. This is becoming a reality with IEEE P7012 Draft Standard for Machine Readable Personal Terms (i.e. “MyTerms”), which has been in the works since 2017 and is now on track for publication early next year.

“MyTerms are contractual agreements about personal privacy that you proffer as the first party, and the company agrees to as the second party. With MyTerms, you don’t “consent” to the company’s privacy policies. They agree to your privacy requirements, which will limit the use of tracking tech to only what you allow. You are not a mere “user” or “client.” You are an independent human being operating with full agency.”

Shelf will ensure this is true with our hybrid, local-first approach and user-set privacy policies. Individuals will be able to decide how their context is shared, and apps must adhere to those terms.

Earning the Right to Context

Context isn’t merely captured - it’s earned. It comes through interfaces that invite authentic sharing from people who trust you with their digital memory. To be the trusted personal context server, you must prove - consistently and in practice - that you serve the individual, not the system. Trust is earned when people see that their history is theirs to interpret, move, and share - and that your infrastructure protects that boundary every time.

The most powerful signals come from vertical surfaces - each with its own unique affordances. ChatGPT will never know your music taste like Spotify or your watching habits like YouTube; each platform captures a different slice of context.

You can’t build Personal Context Infrastructure in a vacuum. You must fit naturally within each person’s digital ecosystem - earning trust and becoming the home for their digital memory.

Why koodos Is Uniquely Positioned

Most companies attempting this space are missing critical pieces:

• AI companies build general-purpose interfaces that miss the high-signal context within specific verticals

• Privacy advocates often build technically sound solutions that real people don’t want to use

• Consumer apps generate engaging experiences and excel at capturing specific types of context, but each operates within its own domain

Consumer apps aren’t the problem: they’re solving real user needs and building valuable context within their verticals. The challenge is that this context remains isolated, preventing the seamless personalization users deserve across their entire digital experience. Anyone building apps today should think deeply about how their products can know the user and act on their behalf - without re-creating the infrastructure from scratch.

With Shelf, we’re building the technical core - context ingestion, inference, and permissioning - that keeps memory under user control. But we’re building more than technology; technology matters only when culture adopts it. Millions already use Shelf to connect their services and keep a digital record of their lives. Shelf isn’t just what you listened to, watched, read or consumed - it’s a portrait of you: your year, your growth, and the culture you belong to. When something appears on more than one person’s Shelf, it becomes culture. For Shelf to feel like a companion rather than inventory management, it must move at the speed of you, your community, and culture at large. We’re building a brand that not only reflects culture - it drives it.

Context Is All You Need

Context is becoming the most valuable currency of the digital age - it will decide which AI services survive in a hyper-personalized future where actions are taken for you. That power must be distributed and placed in service of the individual.

The question isn’t whether Personal Context Infrastructure will be built – it’s who will build it and whose interests it will serve.

We believe it should serve yours, enabling AI that amplifies human potential while preserving and strengthening human agency. Not by competing with the services you love, but by helping them understand you better, on your terms.

Context is all you need. Make sure it belongs to you.

FYI: still figuring out if I prefer personal context infrastructure, or personal memory platform. WDYT?

So here’s how we learned that 500 million tiny files will humble every “proper” engineering approach you can think of.

It started innocently enough. We needed to backfill our data warehouse as part of our new Ingestion Pipeline. Migrate all our historical consumption data from the old S3 bucket format to our shiny new warehouse ready structure. How hard could it be?

We had two buckets to deal with:

• datamover bucket: 50M files (11M actually had data)

• raw-consumptions bucket: 500M files across ~1.7M users

The datamover bucket was going to be our practice round. Get the approach right on 11M files, then scale it up to 500M. Classic engineering, start small, iterate, then scale.

Spoiler alert: sometimes “scale it up” breaks everything you thought you knew.

Act I: The Proper Engineering Approach

“Let’s do this right. Database indexes, sequential processing, proper state management.”

We built what any engineer would build for a data migration:

create table migration_file_index
(
    id              bigserial primary key,
    s3_key          text not null unique,
    platform        text not null,
    job_enqueued_at timestamp with time zone not null,
    file_size       bigint not null,
    processed       boolean default false,
    created_at      timestamp with time zone default now()
);

With proper indexes, of course, we are professionals:

create index idx_migration_sort
    on migration_file_index (platform, job_enqueued_at)
    where (NOT processed);

create index idx_migration_resume
    on migration_file_index (processed, id);

create index idx_migration_platform_stats
    on migration_file_index (platform, processed);

The architecture was clean:

• DatabaseIndexer: Catalog all S3 files that had data into Postgres

• SequentialProcessor: Process files in chronological order

• PartitionBuffer: Smart batching for optimal warehouse file sizes

It felt good. It felt right. This is how you handle a large-scale data migration.

Then we ran it on our practice dataset.

11 million files took one week.

For the datamover migration, a week seemed totally reasonable. Less than a sprint! Exactly what product wants to hear. “Yeah, we’ll have the historical data migrated by next Friday.” Perfect.

Let me repeat that for emphasis: 11 million files. One week.

I remember staring at the progress logs, doing the math in my head. If 11M files = 1 week, then 500M files = ... reaches for calculator ... 45 weeks.

Forty-five weeks. Almost a year. To backfill the raw-consumptions bucket historical data.

That’s when we realized that all our careful database design, our proper indexes, our sequential processing - it was all overhead. We were spending more time managing the migration state than actually migrating data.

Act II: The Cloud-Native Pivot

“Surely AWS can handle what we can’t, right?”

Time for a complete rethink. Why are we trying to manage 500M individual files when AWS has services built exactly for this kind of massive parallel processing?

Enter the Athena approach:

1. Create an external table mapping to our raw-consumptions bucket

2. Let Athena discover and process all 500M files in a single massive query

3. Output everything to a staging bucket in warehouse format

4. Do a controlled cross-bucket migration to production

The promise was beautiful:

• 2-6 hours instead of months

• ~$100-200 in Athena costs vs. massive compute bills

• Leverage cloud services for what they’re actually built for

// Instead of this:
for (const file of 500_000_000_files) {
  await processFileSequentially(file); // 🤮
}

// We’d do this:
CREATE TABLE warehouse_data AS
SELECT 
  platform,
  user_id,
  payload,
  ingested_at,
  date_partition,
  hour_partition
FROM external_consumptions_table;
// Let Athena figure it out in parallel!

We built comprehensive tooling around it:

• Quality validation (95% success rate required)

• Cross-bucket migration with configurable batch sizes

• Progress monitoring and error recovery

• Staging-to-production promotion pipeline

It was elegant. It was cloud-native. It was exactly the kind of solution you’d present at a conference lol.

And Athena choked.

Turns out even AWS services have limits when you point them at half a billion tiny files. Queries would time out consistently. Tuns out 500M files is just too many god damn files!!

The cloud-native dream tried hard, but it died.

Act III: The Big Machine Epiphany

“You know what? Let’s just use a really big machine.”

Sometimes the most sophisticated solution is admitting when sophistication isn’t working.

We spun up an i4i.metal instance:

• 128 vCPUs

• 1TB of RAM

• 30TB of NVMe storage

• 75 Gbps network

The plan was beautifully simple:

1. Download everything to local NVMe storage

2. Process it with maximum parallelism

3. Upload the results

No databases to maintain. No cloud service limits to hit. No complex state management. Just raw compute power and storage bandwidth.

Step 1: Setup the Beast

# Create RAID 0 across all NVMe drives
NVME_DEVICES=$(lsblk -nd -o NAME | grep nvme | grep -v nvme0 | sed ‘s/^/\\/dev\\//’ | tr ‘\\n’ ‘ ‘)
sudo mdadm --create /dev/md0 --level=0 --raid-devices=$NVME_COUNT $NVME_DEVICES
sudo mkfs.ext4 -F /dev/md0
sudo mount /dev/md0 /mnt/nvme

Step 2: Download Everything

# s5cmd with 4096 workers to max out the 75 Gbps network
s5cmd --numworkers 4096 cp “s3://consumptions-raw/*” /mnt/nvme/raw-data/

Enter s5cmd - if you haven’t used it, it’s like AWS CLI’s faster, more parallel cousin. While aws s3 cp might give you decent performance, s5cmd is built specifically for high-throughput S3 operations. It can spin up thousands of concurrent workers and actually saturate high-bandwidth connections.

With our 75 Gbps network and s5cmd cranked up to 4096 workers, we could download at speeds that would make your standard S3 sync weep.

Step 3: The Organize Breakthrough

This is where it got interesting. We needed to transform the data from:

platform/user123/2024-01-15T14:30:15.json
platform/user123/2024-01-15T14:45:22.json  
platform/user456/2024-01-15T14:32:18.json

To warehouse partitions:

platform=spotify/date=2024-01-15/hour=14/part1.gz

Our first approach was to process files sequentially, organizing by timestamp. Still too slow.

Then came the insight: “Wait, we have ~1.7M users here. What if we parallelize by user instead of by file?”

# Instead of processing 500M files sequentially:
for file in $(find . -name “*.json”); do
    organize_by_timestamp $file  # Still 500M operations!
done

# Process ~1.7M users in parallel:
find raw-data/ -maxdepth 2 -type d -name “user_*” | \\
    parallel -j 128 ./[organize-single-user.sh](<http://organize-single-user.sh>) {} /mnt/nvme/partitions

Each worker would:

1. Take a user directory

2. Parse all their timestamp.json files

3. Route each file to the correct date/hour partition

4. Coordinate at the partition level to merge everything

This reframed the problem completely. Instead of “how do we process 500M files?” it became “how do we process 1.7M users in parallel?”

One is hard. The other is Tuesday.

Step 4: Process Partitions

Now each partition could be processed independently:

# Each partition just worries about:
# 1. Collecting all files for that date/hour
# 2. Gzipping them efficiently
# 3. Creating warehouse-format records
find /mnt/nvme/partitions -name “file_list.txt” | \\
    parallel -j 128 --progress ./[process-single-partition.sh](<http://process-single-partition.sh>)

Step 5: Upload and Snowpipe Magic

# Upload to partitioned S3 structure
s5cmd --numworkers 4096 cp /mnt/nvme/output/* s3://raw-extractions-prod/data/

And here’s the beautiful part: once the files land in S3 with the correct partitioned structure (platform=spotify/date=2024-01-15/hour=14/), S3 notifications automatically trigger Snowpipe to ingest them into our data warehouse. No manual intervention. No complex orchestration. Just files appearing in the right place, and the rest happens automatically.

It worked. Fast.

And here’s the kicker: the entire 500M file migration took about a week. The same time it took our initial database approach to handle 11M files.

Let that sink in for a moment. We went from processing 11M files in a week to processing 500M files in a week. That’s a 45x improvement in efficiency, not just from throwing more hardware at the problem, but from completely reframing how we thought about it.

What We Actually Learned

About Scale

500 million files isn’t just “a lot of files.” It’s a scalability problem that broke our initial assumptions and iterations.

The overhead of “doing it right” - indexes, state management, proper abstractions - can exceed the actual work when you hit extreme scale. Sometimes the most elegant solution is the one that avoids the problem entirely.

About Problem Reframing

The breakthrough wasn’t better algorithms or more sophisticated architecture. It was changing the unit of work:

• ❌ “How do we process 500M files efficiently?”

• ✅ “How do we process 1.7M users in parallel?”

Sometimes the bottleneck isn’t compute power, it’s how you think about the problem.

About Cloud Services

AWS services are incredibly powerful, but they have limits. You just have to test them!

“Cloud-native” isn’t automatically better than “local processing with a big machine.” The right tool depends on your specific constraints and scale.

About Engineering Pragmatism

The best architecture is the one that ships. There’s real elegance in recognizing when your careful abstractions are getting in the way and just solving the actual problem.

Sometimes “just use a bigger machine” isn’t lazy engineering, sometimes it’s the sophisticated approach.

The Real Lesson

This isn’t a story about finding the “right” solution. It’s about recognizing when your assumptions break down and being willing to completely change approaches.

• Approach 1 failed because we tried to apply normal database patterns to abnormal scale

• Approach 2 failed because we assumed cloud services could handle anything

• Approach 3 succeeded because we stopped trying to be clever and just solved the problem

The most elegant engineering isn’t always in the architecture - sometimes it’s in knowing when to stop being architectural.

Now our data warehouse hums along, ingesting new files automatically via Snowpipe, and our 500 million historical files are exactly where they need to be. Sometimes brute force is the sophisticated solution.

Want to work on problems like this? We’re hiring widely at Koodos. If this kind of scale, reframing, and pragmatism excites you, drop us a line at team@koodos.com.

As we hurtle toward the development of Artificial General Intelligence (AGI), a system capable of understanding, learning, and performing many of the intellectual and creative tasks a human can, questions of agency become even more pressing.

The Evolving Concept of Agency

In the social sciences, "agency" refers to the locus of action — where power to act resides (Denzin, 2014). Whether it's within individuals, organizations, or even machines, the definition of agency varies across disciplines. Some, like Foucault (1970), argue that agency emerges through discursive practices, while others suggest that machines are merely “objectifications” of human intent (Knoblauch, 2020).

But today, this debate has evolved into something more complex. Can machines possess their own agency, or are they just tools that reflect human will? Social phenomenology suggests that humans often perceive machines as intentional, even though machines merely act based on algorithms and data patterns. This creates an interesting paradox: while we build machines to serve us, they often seem to act autonomously, influencing our behavior in subtle ways (Gunkel, 2020).

“Agency […] is not inherent; it is permitted”

— Henrickson, 2018: 7

As AGI looms closer, this paradox intensifies. If machines gain the ability to perform tasks across a vast array of domains, indistinguishable from human thought, how do we negotiate control? Is AGI simply an extension of our own capabilities, or does it represent a new form of agency, independent from us?

A New Kind of Agency: The Rise of Hybrid Control

Scholars like Latour (2007) and Gunkel (2020) have suggested that agency doesn’t belong solely to humans or machines, but is distributed between them. In this view, agency emerges in the “in-between” — in the relationships and interactions between people and the technology they use. This hybrid agency redefines our understanding of power dynamics, complicating simplistic notions of who (or what) is in control.

Take the rise of AI agents, for example. They’re meant to assist us with various tasks, from scheduling meetings to helping us file our taxes. But while they simplify our lives, they also raise questions about the hidden power dynamics at play. Who really has the agency here? The user? The machine? Or the company that designs and controls the technology?

As AGI progresses, these relationships will only become more entangled. An AGI system that can process immense amounts of personal data, perform complex tasks, and even make independent decisions will redefine the boundaries of agency, forcing us to rethink the roles of humans, machines, and the corporations behind them.

Access Control as a Lever

Consistent and pervasive access controls, embedded in nearly every digital experience, mean that we never truly access content in isolation. Each interaction — whether browsing social media, sharing a photo, or interacting with AI agents — is mediated by layers of authorization and authentication.

These layers create ongoing, often invisible, relationships not just with the digital environments we engage with, but also with the entities that manage, own, and profit from them. This is something we’re tackling head-on at koodos with DataMover and Shelf.

The Privacy Paradox

We often hear that people are concerned about their online privacy — but the truth is that most people don’t.

A part of the reason is that the average person, when faced with the complexity of protecting their data, is overwhelmed. Setting up and managing personal data stores — systems designed to help individuals control and protect their digital information — requires technical know-how. Users must import their data, assign permissions, and continuously monitor for security risks. For most, this process feels like a burden, creating a barrier to entry. This is something we’re exploring more deeply as part of our joint residency with IDEO.

Because of this, consumer value propositions focused purely on ownership and control of personal data struggle to gain traction. While the need for privacy is acknowledged, it lacks the urgency in practice unless it’s tied to a clear utility that impacts daily life.

Utility-Driven Agency

To move beyond the privacy paradox, we need to rethink the role of agency over personal data. Instead of framing it as an abstract right or responsibility, agency must be tied to direct utility — to something that makes life simpler, more enjoyable, or more efficient. When the value of controlling data is tied to real-world benefits, users are more likely to engage.

A great example of this shift is seen in design patterns that focus on user-friendly, intuitive data control. Apple’s approach to data permissioning, for instance, exemplifies this trend. When a user connects a new app to their photo library, they aren’t forced into an all-or-nothing decision. Instead, they can selectively grant access to portions of their library, and they are regularly reminded of this permission, with the option to revoke or adjust it later.

At koodos, we’ve built Shelf — a platform that lets individuals track their digital consumption and curate a public showcase of their taste. By using Shelf, users can aggregate their consumption data in one accessible place. We have been studying the technological shifts around personal data storage and compute — from local-first approaches to private cloud compute — but the challenge isn’t just technical. It's about enabling meaningful engagement with that data, empowering users to reflect on and share their tastes as a means to re-aggregate users’ data in a personal data store that they can control access to over time.

We intend to offer a balance of control and utility that empowers users in meaningful ways, rather than simply overwhelming them with abstract concepts of privacy.

Reclaiming Agency in a Datafied Society

The rapid rise of AI agents, which need access to our private data to perform tasks on our behalf, only amplifies the need for individuals to have centralized control over their data. These AI-driven systems demand a level of access that raises new questions about agency — not just who owns the data, but who has the power to act on it.

Yet, as much as we talk about giving users control, the truth is that even the most sophisticated AI agents don’t function in isolation. They are still shaped and constrained by the companies that build them and the ecosystems in which they operate. As consumers, we are simultaneously actors and acted upon, entangled in a web of relationships that we rarely perceive in full.

As machines become more autonomous, we must grapple with the reality that agency is not a binary, but a spectrum. It’s not simply a question of whether humans or machines are in control; it’s about how we share that control in an increasingly intertwined relationship.

The real challenge, then, is not just about designing systems that give individuals control over their data. It’s about recognizing and navigating the hidden layers of agency that exist between humans, machines, and the corporations that mediate these interactions. The future of agency lies not in choosing between human or machine control, but in understanding how these forces intersect and interact in our everyday lives.

In a world where the lines between human and machine agency are increasingly blurred, we intend to be pioneers in helping users navigate the complexities of a datafied society with tools designed for real utility, control, and empowerment. This is the future we are building — one where agency isn’t just reclaimed, but redefined. We’re excited to be publishing a collaborative paper that discusses the future of the data economy soon.

• Van Dijck, J. (2014). Datafication, Dataism, and Dataveillance: Big Data between Scientific Paradigm and Ideology. Surveillance & Society, 12(2), 197-208.

• Abiteboul, S., & Dowek, G. (2020). The Algorithmic Society: Technology, Power, and Knowledge. Cambridge University Press.

• Hallinan, D., & Striphas, T. (2016). Algorithmic Cultures: Essays on Meaning, Performance, and New Technologies. Routledge.

• Schuilenburg, M., & Peeters, R. (2021). The Algorithmic Society: Power, Knowledge, and the Future of Governance. Routledge.

• Foucault, M. (1970). The Order of Things: An Archaeology of the Human Sciences. Routledge.

• Giddens, A. (1984). The Constitution of Society: Outline of the Theory of Structuration. University of California Press.

• Ricaurte, P. (2022). Algorithmic Cultures and Datafication: A Critical Reflection. New Media & Society, 24(4), 732-748.

• Knoblauch, H. (2020). Communicative Constructions of Reality: Semiotic Mediation and Society. Springer.

• Latour, B. (2007). Reassembling the Social: An Introduction to Actor-Network Theory. Oxford University Press.

• Gunkel, D. J. (2020). The Machine Question: Critical Perspectives on AI, Robots, and Ethics. MIT Press.

• Henrickson, L. (2018). The Entanglements of Agency: Philosophical Reflections on AI, Autonomy, and Human-Machine Relations. The Journal of AI Research, 45(7), 1-20.

• Rammert, W., & Schulz-Schaeffer, I. (2002). Technological Action: Agency and Normativity in Human-Technology Relations. Cambridge University Press.

• Barad, K. (2007). Meeting the Universe Halfway: Quantum Physics and the Entanglement of Matter and Meaning. Duke University Press.

• Coeckelbergh, M., & Gunkel, D. J. (2023). The Philosophy of Artificial Intelligence: Contemporary Debates. Oxford University Press.

On Control, Accountability, and Authority

In an effort to mirror what it’s like building a product and company, I ran an activity where my team had to piece together a puzzle without knowing the intended final image. When building a company, we have to be open to the fact that we don’t know what the final outcome needs to be as the puzzle pieces come together. And yes, even the founder doesn’t really know what the final image is. This led to a dialogue around whether stepping in or standing back was the most effective mode for the CEO to be in, especially considering the unique insights team members brought to specific segments of the puzzle.

Brian Chesky suggests that detail-oriented involvement from the CEO is crucial for maintaining alignment and momentum. He posits that the more proximate the main decision-maker is to the project the better. At the early stages, the founder is domain expert in many aspects of the company, but as the organization expands, the domain expert should ideally be the decision-maker, and often this isn’t the founder. Finding the balance between being in the details and empowering autonomy becomes a sophisticated dance of leadership.

Logically-speaking, I (as the CEO) am ultimately held accountable. And as a leader, I am in a unique position where I could have a lot of control, and could also shift accountability internally. So someone who's directly responsible for a project could technically have no authority (just do what Jad says) and full accountability (if the project goes south, they're ultimately responsible). And so where there’s accountability, there must be agency. And agency requires trust and letting go of control.

A rule of thumb my friend Matt Braley shared is to train the leader do the job to our standards, figure out how to measure the standard, then step away when we see the standard is met 80%+. Naturally, I’ve found that I let go when I trust the person, and I won’t when I don’t. If I don’t trust handing over the reigns, I should really ask myself “can we up-skill the person in a reasonable time so that we’re comfortable?” And if not, then the difficult question should be “why is is this person on the team?”

Despite my ultimate accountability as CEO, or someone's direct responsibility for overseeing progress towards a goal or project, it is essential to establish expectations and foster an environment where everyone takes ownership. That means that no one person is to blame, and that every member on the team is taking 100% responsibility. Everyone accepts their part in both successes and setbacks, openly discusses contributions to challenges, and actively works together to find solutions, fostering a culture of accountability and shared growth. Creating a culture where collective ownership is embraced means that success and failure are shared, fostering an environment of mutual accountability and collaborative growth.

On Whiplash

Navigating a startup is inherently turbulent, with constant learning and shifting priorities. Deciding what takes precedence and for how long is incredibly challenging but critical to progress. This journey demands a blend of urgency and the acceptance of imperfect processes, focusing on agility and on-the-fly problem-solving rather than flawless execution from the outset.

The CEO's actions can rock the boat even more as it navigates the turbulent waters. Underpinning it all is the view that clarity is kindness, and that getting to clarity is the ultimate goal. A few things worth aligning around could include:

• How much time we're "budgeting" for a bet, so that we're committing to a focus for some time, and also so that we're thinking about the scale of the project based on a specific time horizon.

• What part of the discovery process are we currently in? Are we diverging? Exploring different approaches, sketching 100 solutions, running lots of quick tests in different directions, etc. or are we converging: narrowing our focus, testing one thing out and maintaining constraints? 

• The communication flows day-to-day, ensuring that we're collaborating (and not working in silos) as much as possible, and also making room to check in that we're on track for a high impact project and naturally pre-morteming so things just float when there are blockers. Importantly, we should aim to do this in a way that doesn't apply unnecessary pressure or constrains creativity and play.

On Conscious Leadership and ‘Amping It Up’

I find myself toggling between two versions of myself as a leader.

The first: the empathetic, people-first servant leader. A la Concsious Leadership Group.

The second: the driven, results-focused executive that runs towards discomfort and doesn’t accept coasting. A la Frank Slootman’s Amp It Up.

This internal conflict generates both personal anxiety and confusion among my team regarding which version of me they’re dealing with. My goal is to forge an integrated leadership style that harmonizes these aspects, leveraging their strengths to guide my actions and decisions.

On Learning and Anxiety

I tell myself that it’s very natural to be anxious when we're trying to do something very hard, especially when the process requires making a lot of mistakes on the way. Rome wasn’t built in a day, and neither is the process of inverting the internet’s personal data model. Unlocking a retentive consumer product lots of people love is something very very few products can get to — and so I must remind myself that we’ve already achieved more than most startups can dream of.

I had heard many times that startup success is PACE OF LEARNING. But it took me a while to internalize that. A good way to maintain a strong pace of learning is to be in discovery mode – and this applies even at established companies (see: Kodak, or every big tech company that is trying to be bold and take risks again).

One way to reduce anxiety is to shift to a discovery mindset that celebrates learning, not just results. And to trust the process and maintain that learning will lead to results. There’s a lot of inertia against shifting toward a discovery mindset:

• Conditioning: We've been rewarded our entire lives for being right, knowing all the answers, and avoiding mistakes. But perfectionism slows us down when we need to learn by trial and error. We'll suffer endlessly unless we get comfortable with being wrong, not knowing the answers, and making mistakes.

• Incentives: Typically folks get recognized for results over learnings. That’s how we’re typically rewarded through school. This might work well in established businesses where the levers are well understood. But at a startup, we first have to learn what the levers are before we can set the right targets and strive for results. And learning leads to out-sized results.

• The illusion of knowledge: We can't learn things when we believe we already know, so this illusion of knowledge hinders our effort to discover. Recall the bull: beauty is in the simplest outcome (and often we’re stuck amidst the higher complexity phases, and so we actually don’t really know until we get to the simplest form of the bull or can explain a complex concept like a 6 year-old). We often view new information, especially contradictory information, as a threat thanks to our various cognitive biases (including our friends confirmation bias, hindsight bias, consistency bias and sunk cost fallacy).

This requires balancing being a confident visionary with being a curious, skeptical scientist. It's less about feeling confident we have the right answer, and more feeling confident that we can figure it out. When we work from a place of insecurity, we see information as a threat. But new information is the currency of discovery. Great scientists build the confidence to push aside their self-doubt and run toward surprises. To seek discomfort. A discovery mindset means openness, experimentation, focus on users and data, collaboration, bold thinking, and risk-taking.

Subscribe now

Hit reply and let me know if this was interesting or useful, or not. This is an experiment :)

On New York.

I'm frequently asked why koodos labs is building a team in NYC. In essence, this city possesses the highest ambition per capita and the most culture per square foot globally. To establish a generational company that centers culture, these are indispensable elements.

On philosophy in tech

There’s a noticeable lack of deep thinking and introspection in the tech world. It's common to find that foundations are incorrectly laid. However, this isn’t to say we should spend our days in deep thought and avoid shipping regularly to gather feedback and iterate. Scientists also run experiments & talk to people to validate their mental models. They don’t just think.

Unearthing the “secrets” that underpin creative works demands production. The resonance of many profound creative pieces stems from a deep understanding of the human condition; and to get there the creative process necessitates a significant amount of trial and error.

On content discovery

Content discovery is not a burning user need. Yes, there are plenty of problems with content discovery, and yes “we don’t know what to watch” sometimes — but we’ve found that it’s simply not a burning human need and there are plenty of channels to discover content today. People aren’t short on new content. We’ve chosen to frame content discovery as an outcome of our experience, not as a need or value proposition for using our product.

On distribution & the deadly sins.

Attention is the most scarce resource. Platforms, driven by the goal of maximizing time spent and ad views, exploit basic human desires—often summarized as the “deadly sins” of lust, gluttony, greed, sloth, wrath, envy, and pride. This model, when it becomes the foundation of user engagement, instigates a destructive cycle.

In our journey, we’ve anchored on share of identity, instead of share of time. That there’s ongoing utility, joy and insight unlocked from continuous use — that it can be generative and cathartic, without being a time suck or numbing. The idea that Shelf and Self are one letter apart, and that users should be able to see themselves on their Shelf. This is especially important as we shift ownership of one’s identity, and the data that underlies it, to the user.

Nonetheless, the dynamics of distribution are distinct. It revolves around seizing attention amidst the endless demands on it. The deadly sins can make a product resonate, create allure, and quickly occupy mental space. While the sustaining attributes and foundations of a retentive product might not be based on these patterns, it’s evident that effective distribution articulates how a product caters to intrinsic human desires or fears from the onset. It’s possible to align good long-term intentions, avoid gimmickry, and leverage innate human desires for distribution simultaneously.

On the future

I envision a world where diverse tastes are not just tolerated but celebrated; where our similarities unite us and enable us to see beyond the superficial, appreciating each other’s essence. A world where our differences are not viewed as reason for division, but as gateways to uncharted terrains awaiting exploration. We don’t need a second brain, we need a second heart. We need more soul. I imagine a world where our data is unequivocally ours, where we can keep all our stuff in “virtual USBs” that can be connected and disconnected at will, fostering a sense of ownership and empowerment in the infinite expanse of the internet.

Subscribe now

Critics, Analysts, and Tomorrow’s Tech

The narrative in academic spaces like the Berkman Klein Centre tend to have two distinct voices: the critics and the analysts.

Critics, with their incisive observations, lay bare systemic gaps and push for transformative reforms. Their critiques, although usually lagging, capture public opinion and foreshadow regulatory shifts — vital exposure for people building future-oriented technology.

Meanwhile, analysts weave theory and historical insights to paint a picture of what lies ahead. While some ideas born out of academic thought, such as fresh approaches to personal data management, might still be in their nascent stages, they're on the cusp of becoming our reality. It was at the BKC that I had my first exposure to questions around media consumption and self-presentation, the practical limitations of data interoperability, privacy techniques to empower individuals and many of the other themes we touch on at Koodos Labs.

Foreseeing the AI Wave

At the heart of the conference was the question of AI's future trajectory and impact on society. In the shadows of the social media era, there's an urgency to comprehend the potential spirals AI might trace. But what offers hope is the present-day academia and policy-makers' proactive and pragmatic approach, a stride towards anticipated challenges with purpose.

The Quiet Power of Scholarship

To some, academia seems slow and completely disconnected from the reality of building for users and markets. While there’s some truth to that and there’s a very strong need for builders and scholars to meet, it’s precisely the unhurried pace of scholarship, steeped in thought, that grounds us. It’s the time to theorize, discuss, tear down and build again.

Whereas for builders with limited runway and market demands, we're often running against the clock, making it easy to miss alternative approaches or the nuanced implications of our endeavors. When there’s urgency, we tend to retrofit solutions based on what we’ve come to know and it’s very hard to break that pattern. The more access to knowledge we have, the more of an edge we can have as fast-moving builders.

Towards Pragmatic Idealism

I was inspired by Jacinda Arden, New Zealand’s former Prime Minister and now a BKC Fellow. In her efforts to run a country and make policy decisions at speed (like during the Christchurch attacks and COVID), she had limited access to data and had to intuit her way to decisions quickly. She regularly referenced a frame of reference: “pragmatic idealism”: seeking to realize (seemingly unreachable) ideals by taking into account practical considerations. It's a philosophy that resonated deeply with me. When applied to the role of tech innovation, her perspective was that technology magnifies societies existing ills and isn’t a source of our ills. This viewpoint not only shifts the problem's definition but also broadens its responsibility beyond the people building technology — although there’s a need for tech founders to consider societal impact alongside maximizing financial returns — a necessary goal for any venture-backed company.

In closing, as the Berkman Klein Centre marks a quarter-century of intellectual leadership, it serves as a potent reminder of the symbiosis between the tech world, policymakers and academia. Their combined insights foster not only innovation but also responsibility. Whether we're addressing what the future of social media is, the challenges of AI or decentralized technologies, valuing critics and analysts, or pondering the deeper lessons of scholarship, the guiding light remains: to create technology that pushes society forward. It's a journey of continual learning, and I'm grateful to be a part of it.

Subscribe now

In this post we will dive deeper and demystify how apps actually implement authentication. Do it right, and you barely notice it. But do it wrong, and you lock users out or open major security holes.

We'll see how apps authenticate you, from old school passwords to slick new standards like WebAuthn, and we’ll get our hands dirty discussing security and usability trade-offs, encryption, individual sovereignty, and more!

By the end, you'll have a solid grasp of how real-world apps tackle authentication using the methods we've covered.

Let's get started!

Authentication 🪪

The first thing you do when you log into an app is identify yourself.

This might be using some kind of identifier — typically an email address or phone number — or logging in through Google, Facebook, or some other third party service, and using their knowledge of your identity instead.

Taking that identification and ensuring that it is yours is the process of authentication1. Once you’re authenticated, you can use the app!

In the earlier post we explained that authentication can be achieved in one of three different ways. You can demonstrate that you:

• know something

• own something, or

• are something

Only the person you're identifying as would know or own this information, or have that identity.

All modern apps use a technique that boils down to one of these three methods to authenticate you.

So let’s look at each in turn!

Knowledge-based authentication 🧠

Password-based authentication uses the first of these authentication methods — it verifies that you know something only the user you're logging in as would know: their password, pin, passphrase, or any other secret shared between the user and the service.

To prevent attackers being able to access your account in case an app’s authentication data is leaked through a hack, the app doesn’t actually store the password, but a hashed2 version of it.

Roughly speaking, the hash of your password is like an encrypted version of your password, except that given your password’s hash, it’s very hard (near-impossible, if hashed correctly) to retrieve the original password3.

Limitations of passwords

Implementing password-based authentication yourself is fraught with difficulty.

Some of the “fun” challenges you can anticipate are:

• Properly implementing passwords — You need to manage salting, hashing, encryption and then store them securely in your database.

• Password recovery flows — You’ll also need to implement email-based password recovery. But what if your password-recovery emails go to spam? Having angry users contact customer support because they can't log in is not fun.

• Handle "password1@3" — You'll quickly realise that people like to use weak and easy-to-guess passwords.

  This could allow hackers to find a password through a dictionary attack4 — something they can do if they get access to a leaked version of the database, for example — before later gaining access to the running system.

• Block bots — Users also really like to reuse passwords5, so you might have hackers attacking your system because of a data breach on a completely different website. Blocking these hackers when they’re attacking your site from many different IP addresses is a major challenge6.

• Additional friction - There will be quite a few users who won't have the patience to go through the password-recovery flows after forgetting their password. Instead, they just give up and go do something else. Forcing your user to remember something to log-in can add additional friction and increase churn.

• And much more7

Each of these steps is hard and requires scarce engineering resources, which is why authentication providers such as Supabase, Auth0 and Clerk offer managed versions, at a cost.

But most significantly, users may forget their password and never even use your site in the first place!

To summarise: passwords are a poor authentication mechanism. They’re hard to implement, unsafe, and lead to a poor user experience!

So: is there a way to authenticate users without relying on something that only they would know?

I wouldn’t still be writing if there wasn’t, so let’s get into that!

Ownership-based authentication 🔒

Ownership-based authentication is our second authentication method — it involves verifying that you own something that only the user you are attempting to log in as would own, such as their email inbox or phone number.

One-time passwords

For instance, when you log in, an app might send you a one-time password (OTP8) via email. This is typically a random short, server-generated, alphanumeric code like ZXY1, but it can also be a magic link, which is a URL containing the OTP as part of it.

The OTP is a temporary secret that is only shared between the app and the user's email inbox, is valid for that specific interaction, and is invalidated once used.

If you log in using this OTP, then you’ve successfully demonstrated ownership of the email address connected to the user account, and so you must be that user — you’re successfully authenticated!

SMS-based logins can also be used, but SMS are not encrypted and can easily be discovered by anyone on the network by sniffing for text messages9.

SMS-based logins can also be gamed for profit — a big reason why many apps no longer solely rely on SMS for authentication, or even only offer SMS based authentication for premium or paying users.

But worst of all, SMS and emails may take minutes to be delivered, or may never be delivered at all10!

Is there a way for the user and service to coordinate on a shared, time-based secret, without sending it over an insecure, slow and unreliable communication channel?

I haven’t stopped writing, so yes!

Time-based one-time passwords

Authentication apps, such as Google Authenticator, provide time-based OTPs (TOTP11).

These authentication apps are set up well in advance of logging in, typically during the first onboarding session. Instead of using SMS or email to send the OTP to the user, an app on the user’s device generates a TOTP using the current time and a secret shared between the user and the web app.

This is much safer than SMS or email OTP delivery, as it does not rely on the security of external communication channels. The TOTP is generated locally on a device the user possesses and controls, making phishing and interception attacks much more difficult.

Hardware authenticators

Finally, hardware security keys, such as a YubiKey12, are a physical authentication device that a user can carry around with them. These are often used in high-security environments, such as in an enterprise, and can also create OTPs. These are really powerful devices, and we’ll dig into them more soon.

Whilst ownership-based authentication methods are often more secure, easier to use (no need to remember a password!), and easier to implement than passwords, they still require the user to do something cumbersome to authenticate.

This is an opportunity to switch device, get distracted, churn, and never use your app at all!

For this reason, authentication methods that require the user to switch device are used sparingly, only for sensitive operations such as when logging in on a new device for the first time.

So — can we authenticate a user without requiring them to use two devices? Absolutely!

WebAuthn 🕸️

WebAuthn13 is an awesome, new(ish) protocol that allows users to authenticate themselves whilst browsing the web.

Instead of using emails, passwords, YubiKeys, or OTPs, the information needed to authenticate a user is stored in an authenticator — typically the browser or device.

Public-key cryptography

The method used for authentication is public key cryptography. It’s the same technology used in Transport Layer Security (TLS) — a networking protocol which is used for nearly all traffic on the internet. It’s fundamental to security, so let’s take a quick aside to dig in:

The idea behind public-key cryptography, or asymmetric encryption, is pretty simple: each person has two keys — a public and private key. The public key is shared with everyone, and the private key is kept to themselves.

Seems pretty simple so far. But we can now do two magical things:

• Anyone can send the user a message by encrypting it with the user’s public key. Only the owner of the private key — the user — is able to decrypt that message.

• The user can use their private key to sign a message. Anyone with the public key can verify that that user signed the message, and no one other than the user is able to sign it as that user.

Public-key cryptography in WebAuthn

The browser or device, acting as an authenticator, generates a new public/private key pair as the user onboards onto a web service. The private key is kept hidden, potentially stored securely in a hardware security module (HSM14) on the device, and the public key is shared with the web service.

When the user later wishes to log in, the website asks the browser to sign a specific message. The browser or device does so using the private key, the website verifies it is indeed the same user using the previously shared public key, and the user is now authenticated!

https://webauthn.io has an awesome demo of this — try it out!

With WebAuthn, the user doesn’t need to type their username or remember a password and can login with a single click — very impressive!

And excitingly, the prompt to log in with WebAuthn will only appear on the correct site, thereby preventing the phishing attacks that were possible with passwords 🎉.

But logging in with a single click also makes it easy for friends, enemies, and attackers to log in with a single click, if they got close enough to your machine.

So, can we ensure it’s you actually doing the clicking? Surely!

Multi-factor authentication

The HSM can actually prompt the browser for something before it allows access to cryptographic operations, such as generating a public and private key or signing a message. And the browser can prompt the user for this data in turn.

The prompt can ask for a password — another example of knowledge-based authentication — or, even better, ask for biometrics, such as your fingerprint or face. Laptops and mobile phones commonly support fingerprint readers, so this is a fast, easy, and secure15 way for users to authenticate.

Clicking a button to sign in and using your fingerprint is multi-factor authentication, since it requires two forms of authentication for you to log in: ownership of the device with the HSM and your inherent identity and ownership of your fingerprint.

And because you attest to your own identity, this is a form of self-sovereign identity — we’ll talk about this more in the next post.

Cross-device authentication

WebAuthn does pose a problem, though.

Users typically use a website on multiple devices: their phone, tablet, and laptop. It’s easy to share a password between devices — does WebAuthn offer something similar?

Yes! Well, sort of.

You can delegate the authentication to a roaming device — which could be your phone, laptop, or YubiKey — anything that is accessible via bluetooth, NFC, or USB. In turn, that device can sign the message on behalf of the main device.

This is an additional layer of protection, since malware on a site would need to infect an entirely different device than the main one you’re using to steal your private key — neat!

However, this means that we’d always be reliant on that first device for logging into that service, or we would need to set up a private key on each device we want to log in from — hardly a good user experience, and potentially worse than multi-factor login using passwords and OTPs.

Limitations of WebAuthn

WebAuthn isn’t a panacea, and does come with some notable downsides:

• Single point of failure — If you lose a device, you’d lose access to all of your logins that that device stores the private keys for. Recreating those logins would take ages, and would be a massive pain!

  Passkeys16 are an exciting new feature to help plug this gap and to share private keys across devices, allowing for ease-of-access and easy recovery. The fun part is that Google17 and Apple’s iCloud would sync these private keys for you — it’s built into the platform!

  As excited as I am about Passkeys, they are new and controversial1819, and it’s unclear if they’ll achieve mass adoption.

• Inflexibility — Passwords are flexible enough to support most flows that users need20. In contrast, WebAuthn only supports a subset of flows.

  For example, sharing access to a single account (eg to access the bills of an energy provider) with a friend is straight-forward with passwords, but much tricker with WebAuthn.

• Uncertain adoption — Perhaps most important is that users aren’t familiar with the passwordless world.

  Will they onboard? Will they log in again on a new device? Will they understand the user flow? We don’t know yet — and it’s a risk to try out something as innovative as this.

Whilst I really want a passwordless future, and think WebAuthn and Passkeys could be our saviour, there’s one more form of authentication that is widely used and worth learning about.

Identity-based authentication 🙋

The third way we can authenticate is to use something we are — our identity. Biometric authentication is one example of this, and we briefly covered it in the last section.

But we can also implement this by outsourcing identity verification to a third party that users already have an account with — an identity provider21.

When you see “Login with Google” or “Login with Facebook”, that's identity-based authentication in action.

The identity provider tells the app that you are indeed the person you claim to be by providing it with a signed ID token.

When the app you’re signing into gets this token, it uses the signature to verify that the identity provider did indeed create that token — and now you’re fully authenticated!

OpenID Connect

OpenID Connect (OIDC22) is the underlying protocol used for this flow. OIDC is a layer on top of OAuth 2.0 — a protocol used for access delegation23 — that standardises identity data like your name and email address.

Big tech giants like Google, Facebook, Twitter, and Microsoft all offer identity services that implement OIDC. It's easy and convenient for users since they can leverage their existing accounts.

And it saves a ton of headaches for developers since they no longer need to build secure authentication systems. Instead of authenticating a user yourself, just plug into Google, and you're done!

This is an example of federated identity24, where a user’s identity is used across multiple systems — Google, Facebook, and all the apps that rely on these identity providers — instead of just one.

Limitations of identity federation

However, identity federation has a fundamental flaw: the identity provider can revoke access at any time, instantly locking you out of all sites. You don’t own your identity — it’s stored on Google’s servers, not yours.

You may think it’s rare, but it’s surprisingly common2526 to find posts on Hacker News of users being denied access to their Google or Facebook accounts due to one reason or another, and totally unable to access any service that they used Google or Facebook as an identity provider for.

The quest for true user-owned identity is an active area of innovation which I'll cover in an upcoming post! Subscribe now to be updated when it’s released 👇

I would love to hear folks thoughts on this — feel free to comment below or DM me!

Have a great week all — stay tuned for Part 3!

Apple’s about to announce its first foray into mixed reality, an advanced AR/VR headset destined to transport wearers into immersive new realms of connection. It will be the latest attempt by Apple to redefine human connection through media. 

When I started writing this, it wasn’t meant as another speculative piece on how mixed reality might change the world. I remain doubtful people will opt for extended time in clunky and costly headsets. Yet experience warns against betting against Apple in matters of hardware. So, somewhat against initial intention, this is me waxing poetic about the world that could emerge should Apple finally propel us into mixed reality. Perhaps I’m just another fanboy, but if any company can navigate that elusive shoal between human needs and technological possibility, it would be Apple.

Each new medium, from cuneiform to encryption, alters the nature of community and relationship. The telegraph swiftly connected distant hearts; radio gave voices to the voiceless. Social platforms have made connection constant and performative, as if living life connected to an audience.  Virtual and augmented reality promise connection unbound from the corporeal world. No longer confined to sharing thoughts or images after an experience, but sharing experience itself, as lived in a virtual space. What changes when our realities can be shared, not just described? When distance collapses not into text or image, but into a dimensional shift that transports us, Alice in Wonderland-esque, across time and space?

The potential for human connection may be both the means and end of whatever new realities Apple has coded into being. Its devices spread through networks, even as they are networks made manifest, a map of humankind overlaid onto the world. Each heartbeat pulses out a steady beat of intimacy and alienation, and technology quickens the pulse.

New realities bring both promise and surcease, profoundly reshaping the connections between us. How fitting that a company so adept at consecrating intimacy through glass and metal would turn its alchemist’s furnace to distilling human connection into a vaporous stream, fit to flow across the chasms that separate us. The medium is still the message, however rarefied the medium becomes. All else perishes if technology doesn’t join lives and souls. Let’s not forget: connection remains its most fundamental, and lasting, purpose.

With skepticism and anticipation, see you on the other side?

With that context in mind, I’ve been curious about how we form our tastes and what makes us like what we like. My thoughts here are heavily inspired by a number of perspectives on the topic, including the work of Matthew Salganik, John Deighton (one of koodos labs’ advisors), Susan Rogers, Ogi Ogas, Luke Burgis, W. David Marx, and many others.

In today’s piece, I share some of my takeaways from Susan Rogers and Ogi Ogas’ fantastic book, This Is What It Sounds Like — for more, check out their book here.

Before we jump in, a few announcements:

• Will you be at SXSW? We're hosting an event with Crush Ventures. It'll be after-hours at a coffee roastery with a vinyl DJ, tacos and ~some announcements~.

  Request an invite

• The culmination of months of research — I published a high-level framework on progressive decentralization for fellow builders alongside my collaborator, Prof Scott Kominers.

• We coordinated a limited working group on scaling cooperative platform models. The discussions led to a paper that has been in the Top Ten Download list on SSRN for multiple journals since it’s publication.

I identify as…

Our taste profile is formed through biology, experience and happenstance. Your random neural wiring, your exposure to media and culture in the time and place you grew up, and the sheer chance of consuming this piece and not that piece at crucial moments in your life all shaped your taste and influenced the kind of art, music, food, fashion, etc. that you can fall in love with. Your one-of-a-kind taste profile develops and evolves as your internal perceiving, thinking, feeling and social networks learn through experience.

If we were to disregard our broader need for societal acceptance and status — we’ll return to these in a future piece — then our true taste profile at any point in time isn’t something that we choose any more than we choose our height or sexual orientation. The art that tickles our mind isn’t something that we can control. All we can do is be open to learning what this wandering can teach us about our true nature — whether we’re musically attracted to Billie Eilish or Billy Idol. In human relationships and musical relationships, our opportunities for fulfillment start with being honest with ourselves about what we are truly attracted to. In short, the way our mind reacts to art is less something we cultivate, and more something we recognize.

Because each of our brains are wired to experience rewards from different facets of music, we’d be misguided to think that anyone’s taste is superior to anyone else’s.

The neuroscience of liking a song

Our most private dreams and fantasies — the fears, hopes, and longings nestled deep in our psychic core — are all bound together in a neural network in the brain, one associated with our sense of self. One of the best ways to activate the personalized “self network” is by consuming art that resonates with the sweet spots on our taste profile. For example, when we’re immersed in the enjoyment of a favorite record, our mind-wandering network lights up like fireworks.

As our brain is idling, the contents of our reveries — what the psychologist William James labeled “flights of the mind” — contribute to the conscious conception of our personal self. Whenever we daydream or fantasize, our mind drifts to places that are intimate and private, thinking about what we like, what we need, and what we desire. Thus, when we listen to music we love — music that aligns with our sweet spot — we activate the part of our mind that fuels the deepest currents of our identity.

Beyond that, when we’re consuming something, we naturally start to link our personal experience to the experience of the art. Nobel Laureate Eric Kandel explored how when we view a painting, we start to evaluate whether the ideas and feelings it evokes within us match our self-concept. Our memory circuits kick into retrieval mode rather than encoding mode and we automatically “play back” memories of people, places, and events that we associate.

So what do we end up liking? Cultural artifacts that maximize the reward of letting our minds wander — rewards linked to our deepest conception of personal self. And this can be different at different points, for example sometimes we need to access our most-buried emotions, while other times we need to feel our inner dance, warrior or athlete. Sometimes we need words to express our tangled thoughts, other times we’d like to visualize an impossible romance. We turn to our favorite records, videos, poems to take us where we want to go — where we need to go.

Sharing AUX, sharing myself

Gaining insight into a friend’s musical tastes can be an intimate experience that reveals how they see themselves in relation to the world, the value of aesthetic experiences in their lives, or who they want to be when they grow up (or who they wanted to be). In many ways, when we play a song for a friend or send them a recommendation, we aren’t merely sharing the content with them, but we’re letting them in on a part of who we are. This doesn’t just apply to music, but extends to other cultural artifacts.

Interestingly, the identity constructive aspect of culture tends to be more pronounced in our more formative adolescent years. That’s because, neurally, what teens think of themselves is almost identical to their impression of what others think about them. Effective social cognition relies on self-knowledge and self-awareness, and both are under development through adolescence. This is also why lots of social apps, which tend help us temper our view of ourselves, start with a young audience — beyond that group just having more time on their hands.

Beyond a personal resonance, there’s a confluence of external factors for why we like what we like — and also what we say we like. We’ll explore those in a future article.

When we were thinking about ways to incentivize people to re-aggregate all their data in one place, we often found ourselves returning to the metaphor of the shelf, to things tacked up on bedroom walls or record collections. All of these represent stores of our stuff and physical manifestations of who we are, what we are, or how we want those closest to us to perceive us.

Using this framing, of the shelf as a proxy for the self, it’s an interesting coincidence that the words self and shelf are only one letter apart. Because to self-chronicle is to self-construct. The journey of identity construction is intertwined with our active collecting and chronicling of things and ideas. It isn’t that fixed self that we have to actualize or memorialize, it’s the changing and evolving one.

Life online can move so quickly it often feels like a blur. Swipe left, swipe right, swipe up, swipe down. Dings and pings for update notifications. The rabbit holes that algorithms send us down, some of which end up being too good for our own good. And the next thing we know, three hours have passed. In a lot of ways, today’s internet moves at such speed that it puts us in a kind of trance.

Is the ‘metaverse’, being heralded as the next manifestation of our connected lives, going to make things any better? What about the lean-back AI-generated future? One hint may lie in the ‘For You’-page, and the impending FYP-ification of our digital spaces. Everything is being personalized not just down to the level of the individual, but also to the moment at which the individual opens the page. Today’s online experience is dictated by momentum, instead of discrete moments.

What if we could slow things down, so that each digital moment doesn’t merely ‘autoplay’ into the next one? A place where things that truly resonate with us can be captured and serve as extensions of ourselves. Where our online actions are more intentional, more contemplative, and more deliberately non-swipey. What about spaces that feel like “sanctuaries”?

“One of the most necessary corrections to the character of mankind today is a considerable strengthening of the contemplative element in it.” 

— Nietzsche

If we wanted to get big-brainy about it, collecting lifts that moment to the level of consciousness. Of going from just one more nanosecond of life online to something of significance for us. But simply collecting digital things, and moments, isn’t enough.

To produce is to real-ize

In French, the word for producer is ‘réalisateur’ —  or the one who realizes. To produce is to realize. It’s the idea of making something ‘real’, of taking a jumble of ideas and turning them into something understandable and appealing. That’s what good producers do. They make things that don’t yet exist ‘real’. And this applies beyond the context of identity construction. If we think about the process of navigating ideas, we strengthen our understanding of the problem or get clarity on the idea by producing, by shipping, by putting something out there. The act of producing helps you realize the idea and take it from abstract to concrete.

On the advice of a dear mentor, I (Jad) recently took part in a Hoffman Institute weekend retreat and was taken aback by the centering on “Expression” in their process:

In a psychotherapeutic context, you might have heard the advice to sit with that feeling to get through it faster. This is similar. Expressing ourselves — whether that’s by speaking to others, journaling about it, embodying what we’re feeling physically — the act of “expressing”, and therefore “producing”, helps us realize.

On the internet today, consumption is considered the main way to establish one’s particular identity (i.e. we are what we consume) and production is usually ignored in the discourse around identity-forming, since whatever we do outside of work is considered “consumption”. But what happens if we reframe so that more of what we do is framed as “productive” — can some forms of consumption be framed through the lens of “production”?

And the more we realize, the more we become ourselves

The unity of one’s life consists in the coherence of the story one can tell about oneself. 

 — Simon Critchley

As life plays out, we’re constantly re-writing our one-page autobiography. And the story that we tell about ourselves might be different in different contexts, around different audiences or in different points in time.

“Who are you?” said the Caterpillar.

Alice replied, rather shyly, “I — I hardly know, Sir, just at present — at least I know who I was when I got up this morning, but I think I must have been changed several times since then.”

“What do you mean by that?” said the Caterpillar, sternly. “Explain yourself!”

“I can’t explain myself, I’m afraid, Sir,” said Alice, “because I am not myself, you see.”

— Alice in Wonderland, Chapter 5, Advice from a Caterpillar

We’re constantly writing a rough draft of our autobiography. It’s perpetually reshaped by our experiences, but also bounded by where we’re at in our development. The more our story comes together, and the more at peace we are with our story, the more we know ourselves. And, in culture today, we’re in the metamodern phase of rebellion against a fixed state of self — moving beyond the postmodernist view of “performing” as whoever we want to be at any given point in time.

Delineation of the self as part of our story

As we chronicle our lives online, we are forced to distinguish between our role as reader and our role as protagonist. And in that process, we separate the self from the things that influence the self in the story we tell ourselves, and the story we in turn tell the world about ourselves.

We rarely record such a realization of ourselves. Instead we capture bits of our selves, operating in particular moments. But it is this deliberate self-chronicling that becomes self-constructive.

So rather than view our identities as a “nucleus” and our quest to center who we are in one place — to create one single central story, or one ‘profile’ of ourselves — our identity might be better conceived as “distributed”. Distributed across media that’s shaped us. We are a constellation of the things and experiences that define us. Or to use Walt Whitman’s phrase, “we contain multitudes”.

The reluctance of most user-generated content platforms to come to terms with their status as not just a social network but also a personal resource is rooted in this tension. Therefore, any effort to understand the nature and origins of the self is an interpretive effort largely done elsewhere, in parallel perhaps to our life online.

Our understanding of who we are is best served if we view it through a constructivist lens. That way we can see ourselves as the product of the situations in which we operate, the sum of things in which we participate, and put our gaze upon what is hopefully a vast collection of encounters and experiences.

Bye DeFi 🤮

I’m excited(?) the crypto bubble is bursting. Yes, you read that right. Why? Because the money grabs suck and DeFi is exploitative and unsustainable. It taints the whole space. Culling the hyper-financialized projects - including most the NFT art projects -that aren’t solving core needs is AMAZING for the space. So I welcome it with open arms.

Weyl et al said it well:

DeFi was a decreasing returns private property paradigm retrofitted onto increasing returns networks. Built on the premise of trustlessness, DeFi is inherently limited to the realm of wholly transferable private property (e.g., transferable tokens) that mostly bundles rights of use (“usus”), rights to consume or destroy (“abusus”), and rights of profit (“fructus”).

It’s also why I really don’t believe in all the X-to-earn models that pre-suppose the utility of X. It bears repeating, but a play-to-earn game that focuses its efforts on enabling users to profit but that doesn’t get the “play” element right misses the point.

At best, DeFi risks throttling network growth by rent extraction and at worst risks ushering in dystopian surveillance monopolies dominated by “whales'' who harvest and hoover up data in a race-to-the-bottom—much like web2.

So, with all of this, what is “web3” really?

The term “web3” has become the standard label for the next generation of the internet and, more than anything, is a mental model that many developers have agreed to. "Web3" in part represents a shift in consumer demand towards more transparency and consumer control over data. "Code is law" represents a real shift from trust in companies and legal systems to trust in verifiable code, as well as decoupling the individual from the apps they use and centering them in their experience online. This shifts us into “a network of individuals and communities that come together, as emergent properties of each other, co-determining our future…opening this possibility space is a marked improvement over web2’s authoritarianism and DeFi’s anarcho-capitalism”.

Today's authoritarian internet platforms are built on aggregating users (and their data!) to harness network effects – a key source of competitive advantage. BUT how does competitive advantage work in the more open, user-focused ethos of web3 – where, for instance, users can own their digital assets, which are typically created according to interoperable standards on public blockchains (instead of on a company's private servers) - or on non-blockchain based frameworks, such as Spritely, ACDC and Backchannel that rely on data stores tied to local machines rather than global ledgers?

Because a web3 platform's users own their digital assets, they can take them with them when they leave an incumbent platform in favor of a competitor. This introduces new competitive pressures. In web3, new entrants can incentivize power users to move over to them through a strategy called a "vampire attack" that uses an incentive to "suck" users out of another platform (more on how this works, and views of whether it has worked or not here).

Plus web3 platforms run on shared infrastructure layers – typically, public blockchains – which again makes it easier for new entrants to start out: they can innovate on top of existing software, and plug into existing content networks. So while portability of digital assets increases the potential for competition, it also makes the overall dynamics of Web3 less zero-sum – and that means a platform's value creation opportunity can be bigger for all.

The ability to integrate pre-existing digital assets can help resolve the classic "cold-start problem" for new platforms; help existing platforms bootstrap off existing assets (vs. uploading directly) and help source content even for established platforms.

Sharing infrastructure means that web3 apps can focus on building great experiences – driving towards a greater emphasis on platform design as a source of competitive advantage. Even in Web3, design and user insight will continue to differentiate apps and marketplaces. All of this suggests an exciting new era of business strategy is being born, even while drawing on classic strategy frameworks.

Read our full HBR piece here.

Between 2017 and 2021, I taught science lessons to 4th and 5th graders once or twice a week. I’d go into the classroom and teach an hour long lesson around science and the whole point was to help these students feel like scientists. 

The big issue is that rarely do kids see themselves as scientists. Usually, they see science as something other people do. The whole curriculum was built around the premise that everyone is a scientist and that we all do science everyday.

In many ways, web3 is the same way. Many people feel like web3 is too advanced for them, that they aren't ‘the type of person’ who can engage fully with dApps and protocols or contribute meaningfully. 

This is false, and a big part of web3 onboarding, much like my experiences teaching science, is around addressing these misconceptions. The work of onboarding is around increasing confidence, and instilling the fundamental understanding and tools someone needs to thrive in the decentralized web.

Recently, I’ve been focusing my energy on thinking about learning in the context of web3, both how people learn to engage with products and protocols, and how people learn the skills, mindsets, and dispositions they need to thrive in the decentralized web.

As crypto expands exponentially and new people join the space every day, web3 onboarding is an increasingly necessary art and science. At some point, we were all onboarded to web3. Whether it was alone through exploration, guided by a friend or YouTube video, or through a series of blog posts or subreddits. To onboard to web3 is to have the basic technical understanding to participate fully in the crypto ecosystem and the confidence to move forward and grow. To me, onboarding is focused on what it means to become a contributor and participant, not a speculator or investor. 

“Onboarding to web3'' is also extremely vague. It encompses two different, though overlapping, types of experiences: learning how to use a product or platform and learning how to participate in the crypto ecosystem more broadly. The former could be referred to as ‘user onboarding’ and the latter as ‘general onboarding.’ Both are critical when it comes to mass adoption and actually building a better future through web3. From the user perspective, an intuitive and delightful user onboarding will unlock the true potential of consumer crypto products. From the general onboarding perspective, transformative experiences that connect the dots and provide a strong foundation will unlock opportunities for participation, building, and contributing in web3.

A fantastic example of a great user onboarding experience is this video from Friends With Benefits that walks users through how to download MetaMask, buy ETH, swap ETH for FWB on Uniswap, and then link a user’s wallet to their discord account. It’s simple, straightforward, and does a wonderful job of modeling the whole process. 

General onboarding is your cousin Darrell who you haven’t spoken with since 2015 texting you and asking which crypto to buy, and you respond by saying which DAOs to contribute to. You know Darrell is really into Formula 1 and watching reality television, so you suggest he joins F1 DAO and hop into the Mad Realities discord server. After checking things out, Darrell realizes he really vibes with F1 DAO and “wants to buy an NFT membership on secondary.” Darrel doesn’t know what he means when he says this, so you walk him through getting set up with a wallet, buying ETH, and buying a F1 DAO membership on OpenSea. Darrell feels thankful that his cousin understands weird internet stuff, and he’s now onboarded to web3.

As consumer facing crypto products become more intuitive and UX friendly, the complexity of user onboarding decreases. Consider the iPhone, a product with very little onboarding. For the most part, anyone can pick up a new iPhone and begin using it without having to be onboarded. Currently, the remnants of developer-centricity in crypto products remains present, and the novelty of many crypto primitives continues to require intentional and elaborate onboarding designs. Over time, user onboarding will become less complicated as the tools and services become easier to tap into, and general onboarding will become synonymous with upskilling or developing hobbies. 

Below, we’ll explore web3 onboarding from a learning design perspective, and consider best practices for creating the most meaningful and transformative onboarding experiences. 

Planting and Watering Seeds

General onboarding is like planting and watering seeds. Growing seeds is creating structure for users or contributors to become familiar with, exposed to, and integrated within the decentralized web, much like connecting Darrell with opportunities for participating and contributing instead of investing. 

Onboarding is instilling a fundamental literacy and fluency. Blockchains are new languages, both in how they operate, and how we talk about them. When learning a new language, sometimes it’s important to begin with the fundamentals first, and other times it’s most transformative to become immersed, to dive in headfirst. Learning the language of smart contracts, understanding esoteric terms and acronyms, and developing a fluency when it comes to transacting is critical to developing a holistic understanding of the space. The FWB video does an amazing job helping to instill a sense of fluency with crypto transactions. Creating space to learn and integrate these new concepts is critical during general onboarding.  

Beyond language and transactions, general onboarding is connecting the dots. There are two types of dots that are connected during onboarding - dots that connect your personal interests and values to web3 and dots that connect various components of web3 to each other. For Darrell, connecting his interests and values to web3 is him connecting his passion for both golf with F1 DAO and reality television with Mad Realities. Connecting the various components of web3 together is Darrell realizing that NFTs gate access to both initiatives.

Connecting the dots between personal values and interests to various components of web3 is where the magic lies. This is powerful: it shows someone that they have a place here, and that web3 is for them as much as anyone else. For someone interested in environmentalism, connecting them with KlimaDAO or ecodao helps form a deeper connection between them and the wider ecosystem. This is finding one’s niche, finding a digital home, and finding a place to grow. Connecting the elements of web3 together creates coherence, and this coherence leads to confidence. 

More confidence leads to exploration and better question asking. One sign of a successful onboarding experience is the asking of better and better questions. 

Successful onboarding promotes self-efficacy. Without the confidence to continue, newcomers might remain stagnated. Self-efficacy is key to confident exploration, and this is the cumulative product of developing fluency and dots being connected. A sense of confidence is key for long-term success and further navigation, allowing new entrants, like Darrell, to join a discord, make friends, and ask for help.

Generally, if possible, a 1:1 conversation is the best place to start. Ideally, this conversation is between 20 minutes and 1 hour, and between a newcomer with an interest in web3 (your cousin Darrell) and a friend or acquaintance with experience in web3 (you). These conversations allow for fluency to develop, dots to be connected, and curated resources to be shared based on specific circumstances. 1:1 conversations are hard to scale, but when possible, they are often the best place to begin. 

In terms of technical requirements for onboarding, a custodial wallet is extremely important. Wallets are usernames and passports, and getting set up with a custodial wallet is one of the most important first steps. Wallet hygiene is also critical, both around public vs. private keys, and also what it means to operate safely in the crypto space. Unfortunately, scams continue, and given new entrants into the space every day, more and more innovative scams target naïve users. For custodial wallets, MetaMask’s video in addition to the FWB video mentioned above does a great job of explaining the basics.

Once the wallet is set up and the basics are understood, it’s important to know what to do next! Whether this is self-directed exploration, listening to the advice of a friend, or following along a structured guide, successful onboarding is the start of a beautiful conversation. Two phenomenal resources that explore web3 onboarding are the Web3 Starter Pack and Beginners Guide to Blockchain.         

Ultimately, onboarding is a positive sum game and it’s how you get from 0-1. Once someone is onboarded, their journey begins and they are ready to move forward, explore, and onboard others. Network effects are huge, each person onboarded onboards new people and the cycle continues. It’s a beautiful ripple effect. 

We were all onboarded at one point, it’s how we got to where we are, and it’s how we are getting to where we are going. When thinking about onboarding, ask yourself, how did you learn about web3? When did it all click for you? How can you support that in others?

Designing Onboarding

When designing an onboarding experience, I think it’s helpful to use a learner-centered design approach, which could be considered an onboarder-centered design approach. Learner-centered design, “…is the process of building learning experiences by focusing on learner challenges and building fitting solutions by working through an iterative process.” The learning design process (illustrated by Smart Sparrow) involves empathizing with learners, defining goals and objectives, ideating what the learning experience will look like, iteratively prototyping and refining the learning experience, testing it with users, and then reflecting. This process is probably most useful when designing a user onboarding experience for a specific product or platform, and is fairly formal, but provides a great structure for thinking through general onboarding as well.

The most important component of this cycle, in my opinion, is empathizing. When designing an onboarding experience, it is crucial to understand your users, where they are coming from and where they want to go. Do not make assumptions about your users or the person you’re onboarding, and if you have any assumption, validate them. As much as possible, strive for co-designing. One maxim I like is “Never design for your users, always design with your users.” Maintain a reflective and iterative process as much as possible, and always seek the input of your users and those you’re onboarding.

Every onboarding experience is a learning experience and deserves to be treated as such. The best learning experiences are accessible, optimally challenging, and transformative. The best onboarding experiences are the same way.

Key Considerations

Begin with the question: what is the motivation for learning how to use this product or navigate the wider web3 ecosystem? Work backwards from there. Ask yourself what users need to achieve that goal. What fluencies, skills, or competencies are necessary? How can those be learned? And how can I support that learning? 

During any onboarding process, supports need to be established. Whether those supports are an FAQ list, open office hours, or 1:1 mentorship will depend on particular circumstances. Everyone being onboarded needs support, and by understanding your users intimately, you can understand the specific supports they need most.

In addition to support, it’s important to promote just in time learning rather than just in case learning. Just in time learning emphasizes what is currently necessary for the task at hand whereas just in case learning emphasizes what might be necessary down the line. By promoting just in time learning, you support the contextualization of knowledge and skills by allowing onboarders to receive the support they need in the present moment. Over-burdening users with just in case learning leads to lower engagement and understanding, and this is not ideal when onboarding.

It’s also important to not skip corners when onboarding. There is a delicate balance between just in case learning and omitting key information for the sake of simplicity. Much like skipping corners in programming can lead to technical debt, skipping corners in onboarding can lead to ‘learning debt.’ Similar to technical debt, learning debt results from expedited onboarding and becomes exponentially more difficult to remedy. Onboarding creates a foundation for further growth and development. Like building a home, if the foundation is weak, the rest of the building will be unstable, and as you build bigger and bigger, it becomes increasingly more difficult to address underlying instability.

Finally, there is no need to reinvent the wheel. There are so many amazing onboarding and education initiatives who are thinking through this, and my advice would be to consult with similar organizations as you’re designing your onboarding experience.

Happy onboarding!
Maxwell is reachable on Twitter here.

ps. want to shape the future of the consumer internet?

koodos is hiring

also hiring? post to your job here and we’ll share it!

One of the core themes of 2022 will be how decentralized identity will change cyberspace. Below, I will highlight two pieces on the topic:

1. some takeaways from an article I co-wrote for A16Z’s Future on how decentralized identity will allow for derived on-chain reputations that’ll bring about massive opportunities for business and society.

2. I highlight Twitter’s Bluesky project - an effort to develop and drive the adoption of technologies for open and decentralized public conversation - and share some of the core design questions that my peers and I raised for them.

With Decentralized Identity, Your Reputation Travels With You Across Cyberspace

Web3 is based on the premise that each internet user will have a unique internet identifier, like an email address, that can be natively linked to any piece of software and stored on a blockchain. As part of someone’s “decentralized identity,” a portion of a person’s online activity would then be “on chain,” meaning that it would be public and easily searchable via their individual crypto wallet. 

With such decentralized identity — a readable history unique to each person — one’s crypto wallet would function as a sort of profile, similar to Facebook or LinkedIn. But unlike web2 profiles, decentralized identities are backed by hard evidence: a permanent, timestamped record of a person’s accomplishments, contributions, interests, and activities to date. 

If decentralized identity were widely adopted, people would be able to carry their full selves with them as they traverse cyberspace: their affinities and experiences reflected by what they’ve created, contributed to, earned, and owned online, no matter the specific platform. This would bring us closer to how things work in the physical world, where our possessions and reputations are attached to us, rather than to the spaces we occupy; we can take them with us and use them however we please. 

While it might sound far-fetched today, a real opportunity here is in leveraging on-chain reputation for off-chain uses. In the creator economy, for instance, we might see a web3 version of YouTube where the videos don’t belong to the platform, but to the creator. So the creator could bring their videos with them to any online space and reap any value tied to those assets. Likewise, on-chain data could be used to gauge engagement or interest in specific creators or brands. A music artist, for example, could easily identify their top fans based on platform-agnostic, on-chain interactions and reward them with exclusive access or other perks. In the financial realm, decentralized identity could enable an internet-native credit score, which might have applications for issuing loans, performing tenant checks, establishing credit, and beyond. 

continue reading about the infrastructure decentralized identity requires and how decentralized identity will result in downstream opportunities…

Core Design Questions for Twitter’s BlueSky Project: Decentralizing Social Graphs

For a truly decentralized identity, we need a decentralized social graph. Among the promising projects building core infrastructure for decentralized social graphs is Twitter’s Bluesky project. They’ve been engaging deeply with many of the core projects that have explored this question and are trying to build an open standard that Twitter will become a client of. This could mean that Twitter would (once again) open up and allow anyone to operate Twitter clients that compete with its own.

However, this is going to be very hard. Below, I share some important design questions.

1. Is the team designing the protocol and assessing proposals representative of the people who are most hurt/marginalized by these technologies?

2. Are the initial clients of the protocol, beyond Twitter, diverse? This shouldn’t be another case of US-based initial clients and the rest of the world follows.

3. How do we develop the foresight to ensure that data always works in the context of a client’s app?

4. How can BlueSky maintain itself as a non-profit entity and its open-source commitment, without extracting rent from clients?

5. If there was a token, how would BlueSky ensure equitable distribution and counter deflationary effects?

6. How will BlueSky continue to engage with clients to ensure major protocol-level updates are supported to ensure that the protocol continues to be interoperable with as many major social platforms’ standards?

On Discovery & Curation

1. How do we design community-based curation systems with different subjective curation schema?

2. How do we let people make recommendations in a community, collect system-wide signals and protect privacy?

3. How can we enable discovery systems to be community-driven without introducing opportunities for excessive manipulation?

On Moderation

1. How do we ensure that certain bodies with money and influence aren’t able to build “filters” that are perceived to be higher integrity?

2. How will institutions with financial means fair alongside individuals when it comes to these moderation filters?

3. Who decides if a “filter” should be listed -- will Bluesky have to manage the “app store” for moderation filters? Will it list all user-created filters, or exclude (for example) a white-supremacist filter?

check out this “meta-proposal” for Twitter’s Bluesky project that I co-wrote with peers at the Berkman Klein centre…

As a builder in web3, I’ve been concerned by the number of projects that have become overrun by a community of speculators, rather than of actual users. Much of that is due to the way they’ve been bootstrapped around a reputation system that lacks true product-market fit.

At koodos, as we’ve been navigating the idea maze with our community, we’ve been asking ourselves about our reputation system and how we can make contribution rewarding without over-financializing social capital. I previously wrote about how bad incentives ruined the internet and am acutely aware of the misalignment in incentives between users, creators & platforms today. We found that many web3 platforms are grappling with similar questions.

So in the interest of thinking in public (and thanks to Startupy’s GhostKnowledge drop), Scott and I joined forces to develop a generalizable framework. The result:

TL;DR: to make the incentives work, a reputation system should separate social capital from financial capital, particularly if the former offers a clear path to the latter. Scott and I are hosting office hours for anyone who’d like to discuss further! 

To end, here’s one of my favorite quotes about koodos 💚:

"As you pass through the digital world - filled with content, ephemeral and fleeting - koodos is letting people state, ‘I was here, this meant something to me’ and giving them the opportunity to make it a part of who they are online."

Walk into someone’s home, and you’ll find their personal bookshelf. What books? All sorts. Books lent by friends, gifted over the course of a certain chapter of life, ordered off Amazon after being referenced in conversation, picked up spontaneously in the airport duty-free before flights, bought solely for the aesthetic value of their covers…

In this piece, we’ll be using books as a metaphor for ownable cultural artifacts. Let’s get into it.

The bookshelf as a record of who I am...and who I was

My amazing friend Inga Chen wrote:

“What books people buy are stronger signals of what topics are important to people, or perhaps what topics are aspirationally important, important enough to buy a book on it that will take hours to read”....or that they’ll display on their shelf to signal something about themselves.

As I look at the books I’ve accumulated, I’m reminded of how I’ve changed. How my bookshelf changes is quiet, but powerful commentary on what’s happening in my life. A few months ago, I bought a bunch of baking books—like many people, I had a baking phase during covid. Recently, I’ve been really into product design and am amassing a bunch of the canonical books on the topic. In many ways, the bookshelf is an archive of who I am—and who I want to be.

Beyond the selection of books, the way I organize my books is opinionated. I chose to showcase Understanding Media by Marshall McLuhan, but hide Principles by Ray Dalio. I also spent a couple of hours organizing my books based on their color, and every time I add one, I slot into the right place. I care a lot about the aesthetic of my bookshelf, because it’s quite visible - it stands center stage in my living room.

The bookshelf as a source for discovery

When we visit bookstores, we go in with an intention. Sometimes it’s to search for a specific book, to seek inspiration for our next read or maybe we just want to be in the bookstore for the vibes - the aesthetic or what being in that space signals to ourselves or others. When I visit someone’s house, I’ll always look at what’s on their bookshelf to see what they’re reading. This is especially the case for someone I admire or want to get to know better. 

Scanning the bookshelf, I’m on the lookout for a spontaneous discovery. The connection I have with the bookshelf owner provides some level of context and the trust that it’s somewhat a vetted recommendation. I look through the books for a title that catches my eye - maybe I’ll leaf through the pages and sample a few sentences, read the book jacket or the author biography. If something hits, maybe I’ll ask to borrow it or take a note to buy it later. 

The bookshelf as context for connection

There’s something surprisingly intimate about browsing someone’s bookshelf - a public display of what they’re consuming, or looking to consume. When I’m browsing someone’s bookshelf, I’m also on the lookout for books that I’ve read or books that fit my ‘taste’ - and when I find something, it immediately creates common ground, triggers a sense of belonging and connection. It might be even more reason for me to opt to dig deeper into their bookshelf to see what else they’re reading.

Along with discovery, the act of borrowing a book in itself creates a new context through which we can connect. Recommending a book to a friend is one thing, but sharing a copy of a book in which you’ve annotated texts that stand out to you, highlighted key parts of paragraphs—that’s an entirely new dimension for connection. 

Last summer, after falling in love with Sally Rooney’s Conversations with Friends, I mailed my annotated copy to another friend, who then mailed it to another. With each iteration of reads, we kept store of the parts that meant something to us through different colored pens and highlighters, claiming separate parts of the book as our own. It was an intellectual version of the sisterhood of the traveling pants. The copy became a shared collectible that we could use to archive our thoughts, feelings, and emotions, bringing us closer together in our friendship with a new understanding of how we connect with each other, and the messages that resonate with us.

People connect with people, not just content. 

What’s more powerful than the books and the topics they discuss is the author. The effort to source the book, its edition, how early you got it, whether it’s signed by the author, the condition it’s in all serve as some “proof of work” that signal to myself, and the world, the intensity of my fanship. And in all of this, putting out a carrier signal of varying intensity to other fans. 

Take this metaphor of a bookshelf, and apply it to any other space that houses cultural artifacts — or what Julian Weisser has been calling Social Objects. Beyond the books we own, the shirts we wear, the posters we put on the walls of our bedrooms, the souvenirs we pick up — these are all social objects. They showcase what we care about and the communities to which we belong. At their core, social objects have always acted as a shorthand to tell people about who we are, functioning as beacons sending out a signal for like-minded people to find us. On the internet, social objects come in the form of URLs of JPGs, articles, songs, videos.

Pinterest, Goodreads, Spotify and the countless other platforms center discovery and community around creative works - but what’s missing from our digital experience is this aspect of ownership that’s rooted in physicality. We turn to digital platforms as sources of discovery and inspiration, but until now, we haven’t been able to attach our identities to the content we consume. Without public histories that allow for ‘proof of fanship’, there’s no way to track provenance or to verify that you were first to that song, to that artist, to that feeling, commentary, meme. And without true portability, we leave fragments of our digital identity across siloed spaces. 

Carrying our books with us across physical space

When we can’t take our things with us across digital borders, it’s as if the posters on our walls, or the books on our bookshelves are glued in place. It’s as if the clothes we wear leave us as we cross the boundary of one room to enter another.

We imagine a digital world, where our things and our preferences can travel with us across digital borders. Much like how we can pick up our stuff, and take them with us. When we move house, we can pack up our things and transport them with us. And when we move from one room to another, our clothes don’t suddenly leave us. And our preferences surely don’t change, even if how we present ourselves or what part of ourself comes out might change.

As we enter this new era, we’ll see platforms competing for a share of our digital selves, an abstraction of the time we spend online. As more of our lives play out online, we’re spending increasing amounts of time and money developing our digital identities. With that, we’ll see platforms compete less for share of our time, but more for share of our digital identity.

Platforms that become the de facto ‘bookshelf’ for our online lives, where our social objects are placed and are on display, have a huge opportunity in front of them.

koodos has entered the chat. 

Yesterday, I shared an essay I co-authored with Sari Azout on the future of the Internet. People seemed to like it (here, here and here).

Here’s a link to the essay if you haven’t read it yet.

Our hope with the essay was to not only share a conceptual vision for a better Internet (where more people can participate in the upside), but to put our money where our mouth is, so to speak, and come up with an example of how this future might work in action.

To that end, this essay will be one of the first experiments in Attribution+, where we not only cite sources of inspiration, but route economic value to them.

How? That’s where NFTs come in.

We are minting this essay as an NFT and splitting the proceeds of the NFT sale with the group of people who inspired the ideas we share. This is only possible because of our collaboration with Mirror, a startup changing the way writers fund and sustain their work.

The auction was kicked off by Jarrod Dicker’s initial bid and closes in 12 hours.

What are the benefits of owning this NFT?

• Be a proud owner of an experiment that honors the original ideals of the Internet - users and creators, unconstrained by intermediaries, sharing ideas and economic upside

• A powerful signal to the world that you support the “win and help win” worldview, and are, unabashedly, an Internet optimist.

• The prospect of turning a profit by reselling the NFT down the line (Patronage+)

• Own the underlying illustrations we created for the essay.

• We’ll send you prints of the four illustrations (one of a kind because this is the only time we’ll ever print or sell them) which, TBH, I’d hang in my living room - they’re also great dinner conversation starters for the intellectually inclined.

• A 1hr Zoom call with Sari and I to dive deeper into the thesis (anyone who bids over 0.1ETH is invited to participate)

• Early access to koodos and a lifetime Startupy membership (for anyone who bids over 0.5ETH)

Happy Creators

Smaller is Better

A spiritually-aligned Internet

Win and Help Win

Who gets the money?

The funds will be split as follows:

• 50% goes to the authors (Sari and I).

• 30% goes to the thought partners we cite throughout the post. Full list here.

• 10% goes to the people helping us spread the word. If you want to be part of this pool, RT this, give us your take, and add your ETH address. We’ll select the top 10 responses and add them to this pool.   

• 10% goes to Josh Nissenboim, who is behind the illustrations.

Sounds complicated. Why do we need crypto to do this?

• Crypto makes it easy to distribute value across stakeholders in more granular ways. That’s legitimately useful; it automates something that can be onerous in the web2 context. Imagine trying to compensate all of the people we cite in the essay using Venmo or PayPal + a spreadsheet to track? Crypto and NFTs are built for this new world. Writers win by earning money for their work. Thought partners win by earning money for their ideas. Collectors win by owning a piece of Internet history. And you can also win by sharing the essay here. WIN AND HELP WIN.

• An NFT is a simple unique token representing a digital file. And because you can have a limited set of tokens, it introduces scarcity to digital life. Scarcity in the physical world is what drives value - there’s only a limited set of any physical thing, but until now, there was an infinite amount of anything on the web.What this unlocks is powerful: creators can capture value without limiting propagation of their work. When you buy this essay you are not buying this essay, you are buying an entry in a ledger that associates your identity with this essay. A lot of people seem to think that it is completely insane to spend money for something that is freely copyable or accessible. But that misses the point: humans are status-seeking social creatures; we are driven by the need to assert our identity, and there is no better way to do that than through ownership.

• Questions of provenance and authenticity are solved by using the blockchain which is transparent and verifiable. Each NFT has a unique ID and metadata, which you can think of as a digital passport. Moving forward, anytime this essay is distributed on another platform, anyone can "check its passport" and see its entire history, without the need for any third party to intermediate that ownership.

• Perpetual royalties - this is perhaps our favorite advantage of all. The ability for creators to automatically gain a percentage of every sale of their work in the secondary market in perpetuity.

Crypto is hard. How do I bid on this NFT?

Anyone with an Ethereum wallet can participate in the auction here. If you don't have a crypto wallet yet, don’t be intimidated. Consider this your foray into this weird, wonderful world. Follow the steps below:

1.  Turn your regular $ into Ethereum (Cryptocurrency). For most of you the easiest way to do that will be via Coinbase. On iOS you can do this with Apple Pay by using Rainbow Wallet.

2.  Once you’ve setup a Coinbase account or Rainbow Wallet (iOS) purchase some Ethereum

3. On desktop Install a browser-based Ethereum wallet called MetaMask. (Click here for a tutorial on how to install Metamask) On iOS you should already be using Rainbow Wallet.

4. Next send your purchased Ethereum from Coinbase to the Metamask (Click here for a tutorial on how to send Ethereum to your Metamask wallet.) On Rainbow wallet you should already have the Ethereum you purchased with Apple Pay.

5.Once you have ETH in your metamask or Rainbow address, voila, you’re good to go. Click here to bid.

That’s all for now. Let’s see how this goes.

Koodos to you,

Jad

READ THE ESSAY

PLEASE READ THE ESSAY

OBLIGATORY BUTTON TO READ THE ESSAY

The catch? This will be the first ever experiment in Attribution+, where we not only cite sources of inspiration, but route economic value to them.

In practical terms, this means we will be minting this essay as a NFT (click here if you’re wondering what that means), and the proceeds of the NFT sale will be split between the authors, the people that have influenced the essay, and YOU (if you RT this and share your ETH address). It doesn’t get more participatory than that!

‘Jad’s Theory’: the changing definition of talent and wealth distribution online

Every sport that uses ‘mechanical judges’, like stopwatches, instead of human ones are judged solely against an excellence standard. These athletes conform to a taste-free talent vector. We see the entry of a taste vector in Olympic sports that have human judges, such as gymnastics, figure skating and freestyle skiing. To become good at these sports, one must also move towards “connoisseurship”. This requires the development of a “discriminative faculty to judge and express taste according to the aesthetic principles of a consumption domain”1. For a long time, internet content was at the behest of the rigid mechanical judge that is “number of subscribers”, “view count” and others.

To demonstrate the phenomenon of taste distribution more practically, any walk down the supermarket aisles will show that the assortment needed to cater for a diversity of tastes broadens as one moves from say paper towels to soups to wines - products of increasing hedonic value. There is not much dispersion of preferences for paper napkins on any dimension other than price/sturdiness. But that is not true for even the subset of wines that sell in supermarkets. The same applies on the internet.

But what do people end up preferring? Holbrook theorizes that people, as a whole, are much more dispersed in what they prefer than is the ‘expert consensus’. Holbrook’s work draws from Bordieu’s, who famously discussed that those with the highest volume of “cultural capital” dictate what’s deemed “good taste” - and therefore “what’s best”. 

But how do you accumulate the “cultural capital” to dictate “what’s best”? Bordieu, whose work is rooted in the study of the French bourgeoisie, describes how “the working-class ‘aesthetic’ is a dominated aesthetic...obliged to define itself in terms of the dominant aesthetics”. Holbrook argues that in more individualistic contemporary cultures, this gets distorted and we end up with more heterogeneous tastes. 

Winner-take-all outcomes happen in cultural markets under the influence of received expert opinion. But that’s at odds with the internet’s promise to “democratize everything” and to allow us to express our individuality - a culture that flourishes when we can each define what “talent” means to us and that is antagonistic to forced standards and received opinion. 

Thus far, platform algorithms distill our preferences through the ‘mechanical judgements’ of likes, swipes and upvotes and allow an entire social edifice to be built up by iterating them in a recursive fashion. In addition, the “relatedness” of content on platforms today is largely determined by users’ co-viewership patterns2 3, thus the social determinants of consumers’ viewing practices (including one’s cultural capital” is inscribed in the algorithmic output. The resulting ‘clustering’ of content based on “relatedness” into “popular”, “recommended” and “for you” ends up reinforcing clusters of taste4. In that way, platforms have taken on the role of “expert” taste-makers that skew the distribution of taste and that has had a profound effect on what has social and cultural value.

And so, we introduce “Jad’s Theory” (of talent & taste on the Internet). We’re witnessing a renaissance of individualism on the internet - a breakup with institutions and the rise of the individual and, with that, an unbundling of internet communities. For the creator middle class to rise, we need to see higher resolutions of taste preference and a breakup with singular, discriminatory platform algorithms and the opinion of the ‘few’ that arbitrate taste and force today’s dominant aesthetic. With that, individuals can decide on “what’s best” for themselves, allowing for the talent power law to play out across more taste vectors and spreading the opportunity to be perceived as “the best” - and, with that, spreading the opportunity to profit from that.

I’ll leave you with that teaser for now and encourage you to read the rest of the essay.