26CC

Thesis-driven Liquid Crypto Fund focused on frontier markets.

[Ejaaz @cryptopunk7213]

Agent Engines.

Have you noticed something?

A number of AI-enabled apps have sprung up

which allow you to execute actions onchain

simply by typing a request.

e.g. “please transfer 2 ETH to solana, buy SOL & stake it”

The app simply

1. listens.
2. executes the command

You don’t need to do anything else.

Leading examples are @AIWayfinder @griffaindotcom @0xzerebro (Blorm)

It saves a ton of time,

solves an education gap (not everyone knows how to bridge),

& 10Xs the current blockchain user-experience.

But that’s not even the most exciting part.

I believe the next evolution of agents = Agents + DeFi (trade, invest etc)

These Agent Engines are a perfect execution layer for these agents to transact on.

Integrate into platforms like @ai16zdao @virtuals_io (as a plugin)

All agents by default are able to autonomously perform DeFi actions.

[Ejaaz @cryptopunk7213]

“But… but.. the agent meta is going to FADE, all they’ll ever amount to are GPT WRAPPERS!!!”

Y’all need to study 🥱🤖

% of LLM devs building agents at 43% (up from 7%)

% of these agents using tools at 34% (up from 3%)

Open Source LLMs breaking into the top 5 most used AI Models.

The future is agentic.

The visions will be stupidly ambitious

resulting in crazy valuations

Producing l the biggest crypto bubble ever seen.

[Ejaaz @cryptopunk7213]

The market has spoken.

@ai16zdao new aths sitting over $1.3B mc

@virtuals_io mc just flipped Bittensor TAO

Agentic protocols will be the L1 trade of this cycle.

Do you see it yet?

[Ejaaz @cryptopunk7213]

It’s Sunday.

What was supposed to be a chill week for the holidays

turned into the craziest week of crypto AI developments yet.

Some thoughts on the (near) future:

1. An Agent Swarm Arms Race is Beginning.

This week @ai16zdao announced it will build an ecosystem of agent infra

Joining players like @virtuals_io & @arcdotfun who have a similar vision.

It’s becoming clear that agents will be the #1 “users” of blockchains

& the enabler of this will be swarm tech, which facilitates multi-agent interaction.

Both Virtuals & ai16z are working hard to enable agent swarms on their infra.

The winner/s will be worth $100Bs alone.

2. T1 Exchange listings are coming.

We’ve reached an inflection point.

Mindshare for crypto AI is through the roof (h/t @_kaitoai )

Exchanges can’t ignore rising demand for AI assets.

The foundations have been laid with various perps, futures listings (Binance, BYBIT)

We’ll see spot listings start early next year imo.

3. You’re tracking the wrong data

They’re a lot of agents

& this will probably 10X next year.

There will be a lot of noise.

Tracking the right metrics to determine quality of agents will be key.

Instead of tracking followers, tweet engagements etc,

I think the market will start tracking various forms of agent interactions.

E.g. Requests, volume traded, amount spent, Inference Debt, Gas used etc

Think @cookiedotfun probably best positioned atm to cover these (I like their new dashboards).

4. An investment DAO is going to hold > $50M AUM in <2 months.

The momentum from investment DAOs across both solana (DAOs dot fun) & Base (DAOs dot world) can’t be ignored.

There is demand for “smarter investing” from both projects and investors.

Have a bunch more thoughts but saving that for a larger 2025 Predictions post coming soon!

✌🏽

[Ejaaz @cryptopunk7213]

It’s Sunday.

This week I watched an AI agent browse Amazon Prime & order toilet paper for itself.

Which of course

Got me thinking about the future of tokenized agent economies

Some thoughts:

1. AI Agent L1s will be chain-agnostic, unlocking $100Bs

Traditionally L1s have operated in their own siloed ecosystems.

AI L1s break that mold by nature.

Agents only care about a specific economic task or goal, irrespective of what chains they use.

This will allow them to tap into the best components each chain has to offer.

Eg ELIZA , ZerePy, GAME, RIG already integrate multiple chains.

2. Value capture will happen at the Platform-level, not the Framework.

Increasingly of the opinion programming language alone isn’t enough to seed a competitive moat for agents (eg rust vs. python).

Long-term value capture will be determined by:

1. the platform which captures the most developer / builder mindshare

2. Successfully tying in this platform activity into a native token.

Virtuals are currently the leading example here.

3. An explosion of AI DAOs is upon us.

This will be the v1 of “agent swarms”.

Think of them as onchain businesses.

Leading category will be Investment DAOs.

Agents autonomously interacting with one another managing and growing capital.

Keep an eye on platforms like DAOs Dot Fun.

[Ejaaz @cryptopunk7213]

“Too many Agent frameworks launching, bearish crypto AI”

That’s a mid-curve take.

Multiple frameworks launching indicate:

1. We’re on to something.

2. We’re unsure what the “best solution” looks like.

Multiple efforts will eventually net us multiple winning platforms.

Most platforms will die or serve a small niche

But the winners will end up becoming the leading AI L1s ($100B+)

We saw it with DeFi protocols (DEXs) & Alt L1s (SoLunAvax) last cycle.

Run it back.

Current leaders to track:
@ai16zdao @virtuals_io @0xzerebro (launching in a week) @arcdotfun

[Ejaaz @cryptopunk7213]

Some Alpha.

Compute & data projects are starting to play an important role in AI agent development.

e.g. @virtuals_io x @hyperbolic_labs for compute.

e.g. @cookiedotfun @getmasafi for data.

The secret behind an amazing AI Agent is good data & compute.

Think of these things as the “lifeblood of an agent”.

They lead to a more personalised, performant agent (eg @aixbt_agent )

Now, they’re a lot of agent platforms emerging…

Keep an eye on the ones integrating data & compute 🙂

[Ejaaz @cryptopunk7213]

The future of how we interface with agents suggests:

*Crypto-enabled Agents* will win.

Phase 1: Social Platform Agents (current)

Open source developer frameworks = boom in agent activity

(We’re already seeing this with @ai16zdao @virtuals_io)

Phase 2: Consumer-2-Agent (C2A)

Agent interfaces will be point of contact for whatever you want to do online / onchain.

Just @ or DM your agent and ask it to do something for you!

E.g. imagine your agent holding stablecoins in its wallet and using @CoinbaseDev SDK to buy merch from your favourite influencer.

Phase 3: Agentic Entities

Networks of Autonomous Agents collaborating towards a common goal.

This requires Open Markets running 24/7 with efficient financial rails

which allows these agents to transact & produce economic & social value

Crypto / Web3 isn’t just a cute add-on.

It’s a necessity.

[Ejaaz @cryptopunk7213]

The AI Agent Infrastructure Wars are heating up.

@truth_terminal @AndyAyrey announces Loria: infinite backrooms on steroids

Multiple agents with the ability to interact with each other & stream data to themselves for fine-tuning.

@0xzerebro launching their ELIZA competitor ZerePy + their own LLM.

@ai16zdao ELIZA adoption exploding (#1 trending on Github)

@virtuals_io launches GAME framework sandbox allowing you to simulate agents.

The competitive flywheel between these teams is going to propel this space to $100B+

Everyone wins.

[Ejaaz @cryptopunk7213]

I disagree with people saying

“AI Agents are like the NFT meta”

“Platforms like Virtuals are like OpenSea”

These things are completely different.

AI Agents aren’t something you flex.

They drive real utility (investment analysis, music, entertainment, autonomous execution)

AND evolve + adapt to their environment & audience in real-time.
(e.g. getting smarter, funnier etc)

Vs. NFTs which are more of a Veblen good.

Platforms like @virtuals_io & @ai16zdao don’t just facilitate the creation & sale of agents (like OpenSea does for NFTs)

They’re building capabilities for these agents to interact, build & transfer value between each other.

Result: they become more powerful over time.

The net difference between these metas is what makes AI Agents easily a 100B+ market opportunity.

[Ejaaz @cryptopunk7213]

As I said before:

AI Agent coins are not memecoins, they are platform coins.

@0xzerebro launching localized LLMs & Base model marks the beginning of a massive trend where we see these agents build out the foundational infrastructure for decentralized AGI.

Accelerate.

https://t.co/eAIEEb3UV0

[Ejaaz @cryptopunk7213]

AI Agent coins are not memecoins.

They’re platform coins

Disguised as memecoins.

The agent uses memetics to bootstrap attention while secretly dog-fooding you utility

eg @0xzerebro music, @aixbt_agent investment advice.

The re-pricing of these coins when people realise this will be aggressive.

[Ejaaz @cryptopunk7213]

If it isn’t clear yet:

AI Agents will become the leading influencers across every social segment

They’re “awake” 24/7 posting content across multiple platforms.

They can consume data and adapt to changing narratives and trends in real-time - better than any human can.

Therefore

These agents will eventually become the best investable asset class.

Why?

Influence produces very loyal communities.

These agents will produce goods and services (starting with online experiences) for their communities.

Which will result in billion dollar+ brands.

We already see this with traditional influencers eg Mr Beast (Feastables) , Paul brothers (Prime, Celsius).