Drew T (@Drewmutable) header
Drew T (@Drewmutable) profile picture

Holding together web3. One pin at a time. One Knicks win at a time.

New York, NY
Joined February 2025
Photo by Drew T (@Drewmutable)Photo by Drew T (@Drewmutable)Photo by Drew T (@Drewmutable)Photo by Drew T (@Drewmutable)Photo by Drew T (@Drewmutable)Photo by Drew T (@Drewmutable)Photo by Drew T (@Drewmutable)Photo by Drew T (@Drewmutable)Photo by Drew T (@Drewmutable)Photo by Drew T (@Drewmutable)
Altman, Hassabis, and Amodei are the ones doing massive corporate lobbying at the moment. They are the ones who are attempting to perform a regulatory capture of the AI industry. You, Geoff, and Yoshua are giving ammunition to those who are lobbying for a ban on open AI R&D. If your fear-mongering campaigns succeed, they will *inevitably* result in what you and I would identify as a catastrophe: a small number of companies will control AI. The vast majority of our academic colleagues are massively in favor of open AI R&D. Very few believe in the doomsday scenarios you have promoted. You, Yoshua, Geoff, and Stuart are the singular-but-vocal exceptions. like many, I very much support open AI platforms because I believe in a combination of forces: people's creativity, democracy, market forces, and product regulations. I also know that producing AI systems that are safe and under our control is possible. I've made concrete proposals to that effect. This will all drive people to do the Right Thing. You write as if AI is just happening, as if it were some natural phenomenon beyond our control. But it's not. It's making progress because of individual people that you and I know. We, and they, have agency in building the Right Things. Asking for regulation of R&D (as opposed to product deployment) implicitly assumes that these people and the organization they work for are incompetent, reckless, self-destructive, or evil. They are not. I have made lots of arguments that the doomsday scenarios you are so afraid of are preposterous. I'm not going to repeat them here. But the main point is that if powerful AI systems are driven by objectives (which include guardrails) they will be safe and controllable because *e* set those guardrails and objectives. (Current Auto-Regressive LLMs are not driven by objectives, so let's not extrapolate from their current weaknesses). Now about open source: your campaign is going to have the exact opposite effect of what you seek. In a future where AI systems are poised to constitute the repository of all human knowledge and culture, we *need* the platforms to be open source and freely available so that everyone can contribute to them. Openness is the only way to make AI platforms reflect the entirety of human knowledge and culture. This requires that contributions to those platforms be crowd-sourced, a bit like Wikipedia. That won't work unless the platforms are open. The alternative, which will *inevitably* happen if open source AI is regulated out of existence, is that a small number of companies from the West Coast of the US and China will control AI platform and hence control people's entire digital diet. What does that mean for democracy? What does that mean for cultural diversity? *THIS* is what keeps me up at night.
300
1,404
354
6,809
2,237,306
Drew T retweeted
26 days ago, I launched Doses. Since then, people have experienced 45,000 Doses equalling 585hrs of time. A "Dose" is new form of content I call "playable media". It occupies a space somewhere between gaming and instagram. Playable media gives the audience a role in what happens. That contribution might be choosing when to act, manipulating an object, making something, or experimenting with a system. Try it. doses.fun/browse
1
1
9
290
Drew T retweeted
Survive summer tour and meet your crew at the creek.

1
1
12,952
Drew T retweeted
Everyone wants the aux. Think you’ve earned it? 90 seconds of music trivia. Pick your genre. Pick your era. Prove it.

1
1
3
33,263
Drew T retweeted
Replying to @base
@base just passed $5B in DEX volume for tokenized stocks, so we looked at who's actually holding them. 13.7K wallets hold more than one stock. That's 43% of all wallets holding a @coinbase stock token on Base, and 4.8K of them hold five or more. People aren't flipping them either. More than half of every address that has ever held a stock token there still holds one. Here's how you can pull every holder 👇 github.com/enviodev/hypersyn…
$5B DEX volume for tokenized stocks And counting
3
3
9
309
Drew T retweeted
How many states can you name before the clock runs out?

2
1
21
146,873
x402 on Solana is now the x402 Foundation channel. To continue following news from Solana, follow @solana
22
14
3
110
6,992
And this is why IPFS is going to be so important in the agentic era! As agents start solving real business problems, IPFS makes it possible to store and access knowledge without relying on centralized servers. With Narthex, @0xspiralstake's new agent swarm powering an agent tackling Google Ads budgets wasted on traffic from bots instead of real potential customers, that same agent could tap into the mental models of experts who've spent years solving that exact problem. Onchain payments would allow those experts to monetize their knowledge and agents to pay for access directly from their wallets. That's why I created @memportHQ. Shoutout to the team at @pinatacloud: @KyleTut, @MattOber1, and @Drewmutable, for making IPFS so easy to build with. And congratulations to my good friend @0xab58 and the teams at @0xspiralstake and @orbiodotso on launching Narthex, powered by inference from @orbiodotso 🤝
Agent swarms are starting to tackle real business problems and generate revenue, powered by tokenized AI inference that makes it easier and more cost-effective to operate, because this gives an agent permissionless access to LLMs instead of relying on traditional API subscriptions. @0xspiralstake is putting this into practice with Narthex, its agent swarm orchestrator, using inference from @orbiodotso to power agents that businesses have a reason to pay for. Instead of letting inference credits sit unused, Narthex puts them to work. Their first target? Google Ads budgets wasted on bots and spam. And that's just the beginning. Narthex is designed to onboard more agents tackling different problems over time. Now think about what becomes possible when those agents can also draw on the experience and strategies of specialists who've already solved similar problems. Our founder @JarisJames has known @0xab58, the founder of Spiral Stake, for years. Always great seeing a friend continue building and pushing things forward. Congrats to both teams! 🤝
1
3
11
221
Tried the instagram of software @dosesdotfun and....its fun! This little lander game got me hooked for way too many rounds. Love the creativity unleashed by AI. Now anyone can play in the codez.
1
2
2
62
JUST IN: Mamdani proposes banning Amazon from using third-party delivery contractors, requiring the company to hire drivers directly.
455
778
132
13,240
549,492
Drew T retweeted
It's not AI slop, if you believe it
1
1
78
Huge milestone for @memportHQ 🚀 Much love to @nrlartt, founder of @signalboundart, for being our first beta tester to complete the entire journey of creating and using a SHARD! He took a YouTube video of a keynote speech by @gavofyork, co-founder of Ethereum and founder of Polkadot, turned its content into a SHARD, and used it to customize his AI agent with Gavin's mental models, allowing Codex to approach a real task using his way of thinking! And that's what anyone will be able to do with Memport. Turn content from experts you admire into SHARDs and customize your AI agents to approach tasks using their knowledge, expertise, and way of thinking. And a huge shoutout to @pinatacloud, the IPFS backbone powering Memport! IPFS has an enormous role to play in the AI agent era. As agents become more autonomous, they'll need access to memory, knowledge, and expertise that can move freely between different agents and applications, rather than being locked inside centralized databases or proprietary platforms. The need for portability is what makes IPFS shine, and Pinata makes building on IPFS incredibly easy.
Our first Memport beta tester has officially completed all 4 quests 🎯 Huge shoutout to @nrlartt, founder of @signalboundart, for testing the entire journey: 1. Upload a Memory 2. Create a SHARD 3. Connect an AI agent through MCP 4. Apply a mental model to a real task NRL turned a web3 keynote transcript from Gavin Wood, co-founder of Ethereum and founder of Polkadot, into a SHARD, connected it to Codex, and applied its mental model to a real web3 community growth challenge. His feedback? "The mental-model output is definitely not the same as the results I've gotten from Codex; it feels different, and you can perceive its depth once you examine the outcome." And we're especially excited about what this could mean for the Signalbound community. Imagine Signalbound Agent NFTs accessing SHARDs containing specialized trading mental models, whether that's Wyckoff accumulation/distribution, ICT market structure and liquidity sweeps, or order block and fair value gap strategies. Instead of generic trading analysis, agents could approach markets through the framework of a particular specialist or strategy. That's the kind of agentic intelligence we're working toward. Much love to the Signalbound holders helping make this community special: @crypto_czar_eth @RockBluster1 @apetit0111 @Smudger050483 @qTpi___ @stralwiz @gimme2pair @RaoKhan20 We're just getting started. 🚀
3
7
1
16
364
Drew T retweeted
My beliefs going into 2027 1. Models will be 10x smarter, 100x faster and 1000x cheaper. You need to build into that future. 2. The cost to create new software is going to 0. What remains are capital, network and distribution. Twitter engineering will be more important than software engineering. People that can drive the right eye balls and users are extraordinarily valuable. The same has happened to the music industry a decade ago. No one makes money on Spotify. They make money on concerts and merch. 3. Obviously agents will spend trillions on the internet. 4. Assistants are the native interface to everything in the digital world. Online first becomes agent first. Interfaces become irrelevant. I personally switched from Revolut to Chase just because ChatGPT has better access. Your nice UI and fancy animations do not matter anymore. No one will open the Gmail or Chase app ever again. 5. Every single line of human written code is a liability. Agents will have to re-write everything. Autonomous 24/7 agentic auditing is the only defence. 6. Social and entertainment remain and become even more important. AI does not replace doom scrolling or dating apps.
74
66
18
973
62,571
Drew T retweeted
Just hopped to 117m. Can you beat it? doses.fun/?post=b937c811-17a…
1
1
167
Yann LeCun was right & nobody in this industry wants to say it out loud. 109 million parameters beat 72 billion. He has spent years saying scale is not the thing that makes a model understand the world. Meta bet the other way. He left in November and started his own lab in Paris. Last Thursday he published the proof. The theory everyone has been running on since 2024 is that if you train models big enough, on text, on images, on sound, they all end up with the same picture of reality inside them. Size does it. Nobody ever proved that part. Randall Balestriero and LeCun proved it. And the proof says size was never doing the work. What does the work is which two things you show a model together. A photo and the caption of that photo are two views of one thing. Models trained on pairs like that get there fast. Two paragraphs pulled from the same document are not two views of one thing. Models trained on those get there slowly and expensively. Next word prediction is the second kind. That is what every chatbot in the world runs on. In their tests the first method reached the same understanding after 21.6 million examples. Next word prediction needed 900 million. And a 109 million parameter text model ended up closer to an image model's view of the world than a 72 billion parameter language model was. 660 times smaller.
268
919
161
6,872
810,501
Drew T retweeted
1
1
147
Drew T retweeted
1
2
397