
Machine Learning @Latent_Labs, anything from classical statistics to modern machine learning and generative AI, previous @altos_labs
United Kingdom
Joined June 2020
- Tweets379
- Following668
- Followers788
- Likes1.7K
Pinned Tweet
Saying there is a "distribution shift" between train and test data is a great marketing ploy in #machinelearning to make overfitting somehow the data's fault, not the model's or modellers'. 🙃
Sebastian Schmon 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)
Sebastian Schmon retweeted
Today we are happy to share one of our first open research releases: Latent Sparse Attention (LSA), a step towards subquadratic attention that matches accuracy of full attention at ultra-long context.
Sparse attention attends to only a few tokens per query, but it has to find the right ones. At 10M tokens, that search takes 98% of the attention layer's time. LSA makes sparse attention up to 22.7× faster and full attention 72.4x faster.
With @kubicek_ales @caciolai @PengLiangzu @SukritSumant @Ben_Hoov @FraserGreenlee @RMBattleday @radim_urban @thoefler @jcrwhittington
I cannot get over the fact that Max is not the maximum.
We’re renaming our Max plans:
Max 5x → Max
Max 20x → Ultra
Max is still $100/month, and Ultra is still $200. If you’re a subscriber, there’s nothing you need to do.
We’re also removing the 5-hour limit for verified Ultra accounts, starting today with most existing subscribers. Newer accounts keep a temporary limit until they’re verified, which helps us prevent fraud and abuse. You'll get an email when it's removed.
Weekly limits remain unchanged. View our plans here: claude.com/pricing
1
3
124
The extent to which you can just do things now is staggering: I used AI to generate my own RTS, including the possibility to play against an agent by creating a machine readable gaming engine.
I am still wrapping my head around how this will change industries.
1
5
186
I used @claudeai for orchestration, @bfl_ai for designs drafts, @ChatGPT for blender models, @ElevenLabs for voices and sounds. Implemented using @godotengine
2
60
One thing that @OpenAI could do with the massive compute is to brute force re-review every research papers, beginning with maths and physics and continue to stats and ML.
4
204
Unpopular(?) opinion: we don’t need classifiers to detect whether something is LLM generated or not, but rather whether it’s worth reading, i.e. we need to detect slop vs non-slop, not AI vs human.
Those are very correlated but not identical.
4
10
245
Is there a word for a mistake that almost every paper in a field makes, that “everyone knows” is a problem, so nobody can publish the correction? (Rejected as trivial because well known)
4
223
Sebastian Schmon retweeted
uatx.substack.com/p/forget-y… Hmm. I'd say the Enlightenment version of individualism is the historically unique idea that individuals exist not for the sake of their ruler, their class, their ancestors, their gods, or any other such superstition, but for their own sake. Meaning does not require subordination to something "greater than oneself". It is created by the idea that progress is possible and desirable – not by a leap to utopia, but endlessly into the future by undirected problem-solving, carried forward voluntarily by individuals for their own reasons.
Sebastian Schmon retweeted
Not to throw shade but I’m not sure I buy the argument that this is an example of agents “reducing the burden of human expertise” when the prompt used to kick things off is like ten pages long and looks like this. How many humans on earth know all these methods? Maybe 5,000 tops?
I see a lot of confusion around the "achievement of Anthropic" here. It's actually quite simple, this is an example of AI reducing the burden of human expertise.
-Specialized models (e.g. Bindcraft) reduce the burden of access to experiments.
-Reasoning models reduce the burden of access to expert humans.
This is a clear case of the second.
So being an area chair at @NeurIPSConf now means that I have to decide whether the author LLM won the arguments with the reviewers LLMs?












