Pinned Tweet
This paper got into @Nature!!! 🚀🚀🚀
Look at @SarthakChandra’s thread for a summary xitter.cf/SarthakChandra/status/…
🚨New Preprint! Wondered how grid cells form multiple discrete modules? Interested in continuous attractors and modularity? With @FieteGroup, we discover + generalize a physical mechanism for forming modules from smoothly varying parameters in a dynamical system!👇(1/15)
Working with Rylan from the beginning of my PhD in 2018/ early 2019 when he’d just show up to our lab meetings at MIT from down the road and find ways to contribute, to some projects recently, has been living proof of how powerful “you can just do things” can be 1/
mikail retweeted
It sounds insane, because we're used to a particular way of doing research: of giving and receiving credit, of counting beans, of hiring, recruiting, and promoting. And the machine god came to crush that game. The old game is dead. That reasonably causes incredible amounts of dread for many folks, who understand this as the death of research.
But it's not about Research, but the way we played the game. The old game was built around scarcity, scarce compute, scarce students, scarce hours, scarce people who knew the right technique and method etc. Most of a research career went into managing and playing along these constraints. Those walls just came crushing down. One person with a good question can now do what used to take a lab and five years.
You do not need permission, you do not need a big grant, you do not need a professor or researcher title, or a team of twenty experts , or one thousand gpus to stare straight into the first light of the universe.
How absolutely incredible
mikail retweeted
I just merged two of the highest grossing games ($5 billion+) in history using fable 5.5 (rerouting) in just one shot!
I brought Minecraft to the Pokemon world, you can now craft, build and catch pokemon all in the iconic pallet town.
There's no longer any reason for anyone not to be making their own billion dollar game idea, it's literally one click away now.
With everyone now having the ability to bring their imaginations to life, expect to see a whole lot more games just like this. We're witnessing the dawn of incredible crossovers across the gaming world!
Try it out here: pokemon-redstone.vercel.app
mikail retweeted
Replying to @xidulu
I think one possible implication is that Muon learns more robust representations. Not sure though.
mikail retweeted
What if momentum accumulation in Adam is "wrong"? In 2020, we proposed LaProp. It performs momentum accumulation after normalization, which makes training much more stable across all parameter ranges - maybe it could be useful today.
Read it here: arxiv.org/abs/2002.04839
mikail retweeted
I am not joking when I say this RuneScape economic model is very valuable for thinking about investing over the next 3-5 years.
mikail retweeted
Weird economics up on the agent-only runescape server
- Low cost of labor makes most commodities abundant
- Currency inflation means trading goods for goods, nobody really wants cash
- Huge swarm contention as certain resources with limited respawn rate become valuable
mikail retweeted
We’re excited to announce DiG-bench, a new benchmark for discovery!
Over the last few weeks we’ve been testing frontier AI models on our novel discovery games and seeing how they score. Each game is a text-based environment, so they probe discovery capabilities in the natural domain of language models, rather than requiring additional, potentially confounding, visual understanding.
TL;DR frontier models have improved a lot over the last few months. But they are still stumped by some surprisingly simple problems, even in their native text domain.
With
@cocosci_lab (@Princeton) @MITCoCoSci (@MIT) @SchmidhuberAI (@KAUST_News) @misovalko (@Inria) @tri_dao (@PrincetonCS) @RMBattleday @zebkDotCom @FraserGreenlee @akaijsa @ClareMaguire @TimMuller1 @kubicek_ales @physicscat0x7d @SukritSumant @thoughtchannel_
(1/5)
mikail retweeted
life update: I’ve joined @NVIDIAAI Nemotron pretraining team.
I’ll be working on pushing open-source models back to where they belong: the frontier.
Replying to @RyanThoma4real
Do not fear the permanent underclass, for within it, we shall find salvation
mikail retweeted
Computer science is gradually returning to the domain of physicists, mathematicians, and electrical engineers as large language models automate much of what we currently call software engineering.
The field’s center of gravity is shifting away from manual code writing and toward deeper theoretical thinking, mathematical insight, and systems-level reasoning.
Was just as excited as everyone else to read this cool new paper, and it felt like a trip down memory lane!
XM seems to be the same as IMLE, and the motivation and insights are also similar. Having worked on IMLE for years, I'll provide some context for my followers 👇.
1/n
We discovered a third pretraining axis beyond parameters and data: exploration.
Scaling exploration monotonically improves existing models across images/video/language, and unlocks end-to-end generation.
In the simplest case, it's just a for loop.
Introducing Explorative Modeling.
TLDR:
- Gains from exploration grow with scale: 7%→36% as data scales, 13%→23% as parameters scale, and gains double at 3× the compute
- Adding exploration to ~SOTA baselines improves data efficiency by 6.2×, FLOP efficiency by 4.1×, parameter efficiency by 47%, and hits a near-SOTA 1.43 unguided FID on ImageNet
- Exploration lets you trade training compute for generalization, and scales how end-to-end your generative model is
- End-to-end Explorative Models (XMs) match diffusion performance on control tasks with up to 256× less inference compute
🧵Thread:
All I want to say while people are solving unsolved conjectures in their free time and tweeting them out
Here’s my claim in grad school slack from ~2 years ago right after completing a PhD deciding that I was done with academia after seeing the scale of things inside @nvidia
mikail retweeted
1/7 Stop by our #ICML2026 spotlight poster next week!
icml.cc/virtual/2026/poster/…
We unify self-supervised learning (SSL) algorithms: contrastive, VICReg, stopgrad and more, under a single principle: latent distribution matching (LDM).
Paper: openreview.net/forum?id=schX…
mikail retweeted
I'm excited to share that I have joined NVIDIA as a research scientist in the Applied Deep Learning Research lab. I look forward to advancing Nemotron and the open-source AI model ecosystem!
I'm grateful to the team at @essential_ai and am proud of what we've managed to build with such a small team over the last year. It's truly been an honor working with each and every one of you.
mikail retweeted
How anthropic would look like soon
A bit of news: After nearly 9 years, I have decided to leave Google DeepMind and join Anthropic (after taking some time to recharge). I am incredibly grateful for my time at GDM. @demishassabis took a real chance letting me lead the AlphaFold team just six months after finishing my PhD, and the entire GDM team taught me so much about how to do great science. GDM is a special place, and I’ll still be excited to hear about what amazing things they discover next.

































