Alexey Ivanov, CFA (@0xYugiAI) header
Alexey Ivanov, CFA (@0xYugiAI) profile picture

AI Lead & Head of Research · Quant · DeFi

World
Joined July 2022
Photo by Alexey Ivanov, CFA (@0xYugiAI)Photo by Alexey Ivanov, CFA (@0xYugiAI)Photo by Alexey Ivanov, CFA (@0xYugiAI)Photo by Alexey Ivanov, CFA (@0xYugiAI)Photo by Alexey Ivanov, CFA (@0xYugiAI)Photo by Alexey Ivanov, CFA (@0xYugiAI)Photo by Alexey Ivanov, CFA (@0xYugiAI)Photo by Alexey Ivanov, CFA (@0xYugiAI)Photo by Alexey Ivanov, CFA (@0xYugiAI)Photo by Alexey Ivanov, CFA (@0xYugiAI)
Alexey Ivanov, CFA retweeted
AGI might be the most expensive product nobody ordered. Big Tech spends hundreds of billions a year on AI. Most of it runs the models we already have. Why, at this price, nobody needs AGI. Possibly not even the people building it.
1
1
1
3
467
Last month I tried a lot of agent frameworks, skill packs and vibe-coding setups. None of them fit. Not because they're bad. A skill is someone else's frozen decision, made for someone else's work. The good ones are worth studying. What I actually use, I ended up building myself.
2
4
139
Alexey Ivanov, CFA retweeted
AGI is a fascinating dream. But in most debates about it, we underestimate one thing: the economics. The question isn't only whether we can build it. It's who will pay for it.
AGI might be the most expensive product nobody ordered. Big Tech spends hundreds of billions a year on AI. Most of it runs the models we already have. Why, at this price, nobody needs AGI. Possibly not even the people building it.
1
2
128
Alexey Ivanov, CFA retweeted
The number of agents and skills you use defines nothing. You can run dozens of agents and hundreds of skills and still ship complete garbage. Tools don't make the project good. You do.
1
1
2
174
The number of agents and skills you use defines nothing. You can run dozens of agents and hundreds of skills and still ship complete garbage. Tools don't make the project good. You do.
1
1
2
174
A skill is a frozen decision. It works only if you know what good output looks like and wrote it down. Agents won't fix unclear thinking. Hand them a vague task and you get the vagueness back. More of it, faster. Doing the work got cheap. Knowing what the work is didn't.
1
43
AGI is a fascinating dream. But in most debates about it, we underestimate one thing: the economics. The question isn't only whether we can build it. It's who will pay for it.
AGI might be the most expensive product nobody ordered. Big Tech spends hundreds of billions a year on AI. Most of it runs the models we already have. Why, at this price, nobody needs AGI. Possibly not even the people building it.
1
2
128
AGI might be the most expensive product nobody ordered. Big Tech spends hundreds of billions a year on AI. Most of it runs the models we already have. Why, at this price, nobody needs AGI. Possibly not even the people building it.
1
1
1
3
467
Alexey Ivanov, CFA retweeted
It’s interesting how AI and vibe coding are leading people to rediscover the fundamental principles of programming, software development and management.
1
1
175
It’s interesting how AI and vibe coding are leading people to rediscover the fundamental principles of programming, software development and management.
1
1
175
The argument that CoT monitoring is already weak is correct and it does not automatically license moving reasoning off the text surface. Both things can be true. To their credit they seem to know that.
meet @mostik_ai! what happens when you put 12 PhDs in one room for four months? first place on the ARC-AGI leaderboard, which I can't say much about while the competition is still running. and this, which I can. everyone's arguing about whether open models will catch up to frontier models. we think it's the wrong question. here's the one we pose: why does a frontier model have to generate your answer at all, when the only thing you need from it is the reasoning? we do this by enabling models to communicate in latent space. through our protocol, hidden states pass straight from a frontier model into a small one running on your infrastructure -- no text between them, and neither model is fine-tuned. two models from different families, sharing reasoning, both left untouched. how do we know it works? we tested it on a setup where a 753B model reads the problem, and a 4B edge-class model writes the answer. with this approach, we get results 80% as accurate as the frontier model, but at 20x faster performance. we're committed to preventing frontier model lock-in and are already partnering with inference providers to accelerate open-weight adoption. we've done this between 15 of us, in four months, 12 PhDs and a Fields medalist, backed by @generalcatalyst WIRED has the first external account of the company and the work: wired.com/story/russian-star… full writeup, the setup, and all the numbers: mostik.ai/read-more
2
1
8
434
Right now we don't need to reinvent payments or build new infrastructure from zero. All that's needed is an upgrade under the hood one where users don't even notice they're using stablecoins instead of SWIFT.
1
4
235
Alexey Ivanov, CFA retweeted
Plenty of non-USD stablecoins exist, but USD-denominated ones, USDT and USDC, still dominate the stablecoin market. nitter.cf/0xYugiAI/status/208281…
EUR stablecoins doubled in 7 months. BRLA went from zero to $400M/month. And 99% of the stablecoin market is still the dollar. Both are true. Here's why non-dollar stablecoins aren't becoming an alternative — they're becoming something narrower.
1
2
278
Plenty of non-USD stablecoins exist, but USD-denominated ones, USDT and USDC, still dominate the stablecoin market. nitter.cf/0xYugiAI/status/208281…
EUR stablecoins doubled in 7 months. BRLA went from zero to $400M/month. And 99% of the stablecoin market is still the dollar. Both are true. Here's why non-dollar stablecoins aren't becoming an alternative — they're becoming something narrower.
1
2
278
Alexey Ivanov, CFA retweeted
EUR stablecoins doubled in 7 months. BRLA went from zero to $400M/month. And 99% of the stablecoin market is still the dollar. Both are true. Here's why non-dollar stablecoins aren't becoming an alternative — they're becoming something narrower.
1
1
1
5
881
EUR stablecoins doubled in 7 months. BRLA went from zero to $400M/month. And 99% of the stablecoin market is still the dollar.
EUR stablecoins doubled in 7 months. BRLA went from zero to $400M/month. And 99% of the stablecoin market is still the dollar. Both are true. Here's why non-dollar stablecoins aren't becoming an alternative — they're becoming something narrower.
2
4
295
EUR stablecoins doubled in 7 months. BRLA went from zero to $400M/month. And 99% of the stablecoin market is still the dollar. Both are true. Here's why non-dollar stablecoins aren't becoming an alternative — they're becoming something narrower.
1
1
1
5
881
Alexey Ivanov, CFA retweeted
The crypto card situation feels a lot like DeFi summer, when the space got flooded with copy-paste apps and straight-up forks. Same thing now. So many crypto cards doing basically the same thing, some are just clones, and new ones drop every day. But here's my question: do we really need this many? I use @ether_fi or @KASTxyz, open them almost anywhere, spend globally. So why would I need five more that do the exact same thing? Sure, some pitch themselves as "built for local markets" or some niche use case. Honestly though, a decent general card already covers most of that. My bet: a lot of these projects quietly die off. If you're not solving an actual problem or doing something genuinely different, nobody needs you, and you disappear the moment investor money dries up.
2
1
3
264
The crypto card situation feels a lot like DeFi summer, when the space got flooded with copy-paste apps and straight-up forks. Same thing now. So many crypto cards doing basically the same thing, some are just clones, and new ones drop every day. But here's my question: do we really need this many? I use @ether_fi or @KASTxyz, open them almost anywhere, spend globally. So why would I need five more that do the exact same thing? Sure, some pitch themselves as "built for local markets" or some niche use case. Honestly though, a decent general card already covers most of that. My bet: a lot of these projects quietly die off. If you're not solving an actual problem or doing something genuinely different, nobody needs you, and you disappear the moment investor money dries up.
2
1
3
264
Crypto cards are a solid bridge between crypto and real life. But there are too many now. Users don't need a dozen cards, they need a few that actually solve a problem. Just another me-too card won't survive.
1
160