Energy, Taste, and Intelligence Cofo @Basethesislabs, e/acc @GenesisAILabs
Earth
Joined September 2019
- Tweets773
- Following1.9K
- Followers1.2K
- Likes20.5K
Sid retweeted
Dear Prime Minister Ambani,
Please accept my humble apologies for not realizing that you are the real boss of India.
Naturally, you would prefer to maintain your monopolistic exploitation of the great people of India, but would you nonetheless consider allowing Starlink to compete?
There are many areas of India, as there are in all countries, with no Internet access, which denies children the opportunity for education and for small businesses the opportunity to sell their goods to a global market. Starlink would change their lives for the better.
Thank you 🙏
Sid retweeted
really well written blog on the emerging shift from optimization-limited to rollout-limited frontier RL. what stood out to me was the observation that as reasoning trajectories become longer and more heterogeneous GPU utilization and learner-generator synchronization increasingly dominate training efficiency. one underexplored question is whether capability jumps themselves can be detected through staleness statistics. if old trajectories suddenly become much less useful for learning that may indicate the model has discovered a qualitatively new reasoning strategy. in that sense replay-buffer value decay could become an observable signal of phase transitions in capability acquisition. if you're into async-RL/large-scale training systems this is a really worthwhile read.
New blog! Is frontier asynchronous RL solved?
The blog covers Async RL theory and infrastructure, surveying 8 open-weight frontier labs for the algorithmic techniques and systems fixes to handle train-inference mismatch. Also answered: why do current methods still fail at high policy lag? Which methods scale with horizon and compute?
Sid retweeted
The open-source protein ML space just got a massive upgrade. Phenomenal work by @anindyadeeps and @try_litefold on dropping the biggest protein data collection on Hugging Face
We have released the biggest protein data collection on Hugging Face, guys!
We have been working on this for more than 3 weeks now, starting from curating the raw data, doing a lot of filtering, splitting the datasets, sharding them, and doing a lot of analysis. Everything is summarized in our recent blog post.
Sid retweeted
The bitter lesson in 26 words:
Don’t be distracted by human knowledge, as AI has been historically.
Instead focus on methods for creating knowledge that scale with computation, like search and learning.
Sid retweeted
🚨Excited to announce Agent-BRACE!
LLM agents in long-horizon POMDPs either blow up their context with raw history or summarize it, discarding uncertainty by collapsing belief into a point estimate. Agent-BRACE decouples the agent into belief state + policy models, jointly trained via RL.
Key takeaways:
1️⃣ 🎯The belief state model produces a structured approximation of the belief distribution as a set of atomic natural-language claims with ordinal verbalized certainty labels ranging from certain to unknown. The policy conditions on this compact belief rather than the full history.
2️⃣ 📈 Outperforms strong RL baselines on long-horizon partially observable embodied language environments while maintaining a near-constant context window independent of episode length.
3️⃣ 🔄 The learned belief becomes increasingly calibrated as evidence accumulates, and epistemic belief decreases over time: the proportion of claims that the agent has the strongest level of belief in grows from 21% → 52% over an episode.
👇🧵
Sid retweeted
Amazing talk by @DrJimFan -- exciting times for Physical AGI!
TL;DR
VLA architecture is parameter-misallocated toward language and should be replaced by World Action Models
→ pretrained video diffusion models that jointly predict future world states and robot actions, instantiated by Dream Zero (a 14B model running real-time control at 7Hz with 2× generalization gains over VLAs, @SeonghyeonYe ).
arxiv.org/pdf/2602.15922
His central data claim is that egocentric human video is the FSD-equivalent ambient data flywheel for robotics, and EgoScale (@ruijie_zheng12) demonstrates a near-perfect log-linear scaling law (ℒ(N) = a − b log N, R² = 0.9983) between 1K and 20K hours of pretraining data and downstream dexterity performance.
arxiv.org/pdf/2602.16710
His central environment claim is that classical physics simulators will be replaced by neural simulators, and Dream Dojo (@ShenyuanGao) demonstrates this with 44K hours of human video pretraining, 10 FPS real-time interaction, and Pearson r = 0.995 policy-evaluation fidelity.
arxiv.org/pdf/2602.06949
Significant gaps in the talk:
- it does not address runtime semantics (skill installation, behavior consistency, run-update separation)
- it does not address the model-exploitation failure mode of training policies against learned simulators or learned rewards.
My running notes w/ Opus 4.7 👇
docs.google.com/document/d/e…
@NVIDIAAI
I promise this will be the best 20 min you spend today! Robotics: Endgame, the sequel to my last year's Sequoia AI Ascent talk, "Physical Turing Test". I laid out the roadmap for solving Physical AGI as a simple parallel to the LLM success story. Be a good scientist, copy homework ;)
And stay till the end, more easter eggs and predictions for your polymarket!
00:30 DGX-1 origin story at OpenAI, I was there in 2016 signing with Jensen and Elon. Heading to the Computer History Museum!
01:42 The Great Parallel
03:31 Robotics, the Endgame
03:39 Why VLAs fall short
04:32 Video world models as the 2nd pretraining paradigm
06:09 World Action Models (WAM)
07:46 Strategies for robot data collection and the FSD equivalent to physical data flywheel for robot manipulation
11:06 EgoScale and the Dexterity Scaling Law we discovered recently
14:00 Physical RL: bridging the last mile
15:39 DreamDojo: an end-to-end neural physics engine for scaling RL in silico
17:00 Civilizational Technology Tree and my predictions for the near future. Spoiler: it's closer than you think.
Thanks to my friends at Sequoia for inviting me back to AI Ascent this year! I had a blast! Last year's talk is attached in the thread if you missed it.
Sid retweeted
Vibe Coding is not vibing? Agents perform <60% on SWE-WebDevBench 👀.
Code-level benchmarks (HumanEval, SWE-bench, FeatBench) take the specification as given and grade the patch. Vibe coding inverts this: the user gives natural-language intent, the platform must do PM, engineering, and ops in one pipeline.
We evaluate across three orthogonal axes:
➡️ Interaction Mode (ACR=App Creation Request, vs AMR=App Modification Request),
➡️Agency Angle (PM × Engineering × Ops),
➡️Complexity Tier (T4 multi-role SaaS, T5 AI-native).
Introducing SWE-WebDevBench: a comprehensive eval framework to assess AI coding platforms as virtual software development agencies, covering not just the middle step of coding, but the entire software lifecycle: Requirements gathering, planning, deployment and change management.
@QwikBuild, @emergentlabs, @Lovable, @v0
see webdevbench.com 🧵
Sid retweeted
Coding benchmarks like SWE-Bench and the latest one ProgramBench are useful, but can AI coding platforms like @qwikbuild, @replit and @lovable actually build and maintain real-world web applications?
Introducing SWE-WebDevBench: a comprehensive eval framework to assess AI coding platforms as virtual software development agencies, covering not just the middle step of coding, but the entire software lifecycle: Requirements gathering, planning, deployment and change management.
Sid retweeted
SimScale
[CVPR 2026 Oral] Learning to Drive via Real-World Simulation at Scale
🏗️ A scalable simulation pipepline that synthesizes diverse and high-fidelity reactive driving scenarios with pseudo-expert demonstrations.
🚀 An effective sim-real co-training strategy that improves robustness and generalization synergistically across various end-to-end planners.
🔬 A comprehensive recipe that reveals crucial insights into the underlying scaling properties of sim-real learning systems for end-to-end autonomy.
Sid retweeted
worked on self compaction at @lossfunk with @sidgraph earlier this year. wanted to see if compact as a tool actually works. let the model decide when to compress its own context, see how far that gets you.
a lot of learnings and failures along the way. wrote it all up here!! 👇👇
[co-written with claude opus 4.7.]
Sid retweeted
hey, if you're in blr you should drop by. we'll be talking rl, post-training and more!! should be a fun one, don't miss it.
We’re hosting an invitation only gathering of researchers primarily in AI (but not limited to) over dinner in Indiranagar BLR!
Just thoughtful conversations over good food, about seminal papers, emerging fields, research of your group/ recent papers you’ve published 🫶
If you are interested/ doing research in:
- Coding Agents
- RL (Synthetic RL envs, Rl post training, Alignment, On policy distillation)
- World Models
- Neural Computers / Self adaptive neural systems
We would love to have you over dinner!
PS: Amazing venue, chill vibes, good food and most intellect people from BLR all together
Hosted by Me and @thesisofsarthak from @Basethesislabs <3!
Register through link below👇
luma.com/zndszwdp
Sid retweeted
I have been talking to Base Thesis lab folks for a while, super high agency, and really cool insights, come along!!
We’re hosting an invitation only gathering of researchers primarily in AI (but not limited to) over dinner in Indiranagar BLR!
Just thoughtful conversations over good food, about seminal papers, emerging fields, research of your group/ recent papers you’ve published 🫶
If you are interested/ doing research in:
- Coding Agents
- RL (Synthetic RL envs, Rl post training, Alignment, On policy distillation)
- World Models
- Neural Computers / Self adaptive neural systems
We would love to have you over dinner!
PS: Amazing venue, chill vibes, good food and most intellect people from BLR all together
Hosted by Me and @thesisofsarthak from @Basethesislabs <3!
Register through link below👇
luma.com/zndszwdp
Sid retweeted
We don't even design binders!!
But we @BioMandrake just won @adaptyvbio × @gembioworkshop's RBX1 design competition - 1 Strong binder out of 322 tested, selected from 12,000+ submissions.
Couldn't make it to @iclr_conf in Rio. Here's how we did it 👇
open.substack.com/pub/mandra…
Sid retweeted
Think of proteins like IRK as biological power switches for our cells. The DFG motif is actually located in the orthosteric part of the protein, but it is the key that unlocks the allosteric site. We can leverage the DFG-out conformation and design highly specific inhibitors.
Sid retweeted
> Be David Silver
> Study at University of Cambridge where you meet @demishassabis
> Do PHD in reinforcement learning in 2004
> Co-found video games company
> Become a lecturer at UCL
> Record a series of reinforcement learning lectures series on YouTube
> Join Google DeepMind early in 2013 as a core RL researcher
> Develop AlphaGo and beat a world champion at Go
> Develop AlphaZero, mastering chess, shogi, and Go from scratch without human data
> This fundamentally changes the way the world sees AI
> Contribute to mulitple other projects
> Win numerous awards
> Become one of the most cited AI researchers on the planet
> Reject the view that LLMs will ever develop AGI
> Quietly incorporate Ineffable Intelligence in London in 2025
> Step away from DeepMind in January 2026 to lead the startup full-time
> Raise $1.1bn at a $5.1bn valuation
> Promise to give all money made from the company to charity
This guy is building AI that can produce genuinely new knowledge.
If successful it could destroy every LLM provider there is.
BREAKING: @IneffableLabs has raised an INSANE $1.1bn Seed round - the UK's and Europe's largest EVER 🇬🇧 🇪🇺
The round values the company at $5.1bn.
The company has been founded by David Silver - one of the earliest researchers at DeepMind who was responsible for reinforcement learning.
This is a way of training AI models from experience rather than historical data (i.e. they learn by trial and error as opposed to learning from historical data).
David was involved in some groundbreaking projects at @GoogleDeepMind, including those responsible for defeating the world champion in Go.
Today the company has announced a $1.1bn seed led with backing of the following investors:
> @sequoia: @Alfred_Lin, @sonyatweetybird
> @lightspeedvp: @ravi_lsvp, @ravirajjain
> @EvanticCapital: @MattEvantic
> @UKSovereignAI: Joséphine Kant, @Jameswise
> @Indexportfolios: @dannyrimer
The aim of the company is to create an AI superlearner that can discover knowledge from its own experience. This new capability will be driven by the world’s most powerful reinforcement learning algorithms.
If this goes well this will become more powerful than any LLM created. And it's being built here in Europe.
LETS GO
Sid retweeted
Help us make Biotech cool again.
We're living through biology's most interesting decade - gene editors, enzymes designed atom by atom, proteins folded on laptops - and the industry still dresses like a pharma brochure from 2008.
@MandrakeBio is looking for a design partner to do something about that. DMs open.
Sid retweeted
🚨Hiring Member of Technical Staff at @Basethesislabs
We are a frontier lab focused on democratising the way humans interact with technology to maximise their potential.
If you are interested in building world models, continual learning, and RL environments, apply for our research roles.
If you are a serious builder with proven impact in deploying production-grade agentic systems, apply for our engineering roles.
PS: We provide dedicated Claude code accounts, unlimited console play hours + everything else you get in your regular jobs!
binary.so/YBa7Dak
