Joined May 2008
- Tweets 4,819
- Following 1,128
- Followers 1,086
- Likes 6,165
411 Photos and videos
Pinned Tweet
1 Sep 2025
1/ Our paper “Protein Language Models Expose Viral Immune Mimicry” is now published in Viruses!
We show that protein Language Models can identify viral proteins, and those that fool our immune system.
2
1
17
1,588
Dan Ofer (✈️ ICML-26 Was @Worldcon) retweeted
Jun 12
Must read by all those who think that a real neuron is equivalent to a single weight in a deep learning model.
Jun 11
What can a neuron compute?
Real biological neurons are complex, but how capable are they?
Using a new method, we found that a single cortical neuron can classify cats vs dogs, recognize spoken words, and solve 10-bit parity, all tasks thought to require entire networks. (1/15)
2
13
93
10,950
Dan Ofer (✈️ ICML-26 Was @Worldcon) retweeted
Jun 11
Excellent thread. We always use CLR transformation in metagenomics, so I was confused why it wasn't used in single-cell genomics
Apparently the "CLR" function in the most common single-cell analysis tool failed all the benchmarks because it was implemented incorrectly 😶
Jun 10
Arguably the most boring step in genomics is the first one: normalization. Settled science. Scale log. Move on.
Except that here's been a huge blind spot in the field. And it matters for AIxBio. A 🧵about what I think may be one of the most important papers I've written. 1/
6
11
97
14,452
Dan Ofer (✈️ ICML-26 Was @Worldcon) retweeted
Jun 2
This is the workflow I want beginners to copy: do not ask the agent to finish, ask it to prove. Plan, change, test, summarize what failed, then fix. That feedback loop is where Claude Code becomes a teammate instead of a guessing machine.
1
4
2,757
I'll be at #ICML 2026 in Korea at the workshop! Showing 2 of our works in LLMs for Bio/FM4LS Workshop :D
Ping me !
[ICML] Int'l Conference on Machine Learning
4
530
RT @DrJacobsRad: The world is about to get a master class in antisemitism.
Israel's minister Ben Gvir's behavior towards the flotilla acti…
2,918
Well, i'm an idiot. I read as - , in the delta of the ablation changing my loss. Dohhhhh
47
#Autoresearch on #proteins - amazing what you can do with a gpu and a few nights on the SOTA for language models.
(Now I just need more gpus)
Viva @karpathy
#bioinformatics
1
5
377
Dan Ofer (✈️ ICML-26 Was @Worldcon) retweeted
Pancreatic cancer is one of the most dire diagnoses in medicine with few available treatments. Until now, thanks to university research, including @UCSF scientists, and federal investment in science research. Read about this huge breakthrough via @nytimes nyti.ms/4wfziXs
2
33
123
42,490
Dan Ofer (✈️ ICML-26 Was @Worldcon) retweeted
May 8
May 8
Exactly - biology is a fundamentally different domain than text and scaling laws do not apply cleanly
~All the tasks you want an LLM to do are contained in the text data itself. For biology, NONE of the tasks you want the model to do are contained in the sequence data itself.
4
10
169
13,066
Academics Need to Wake Up on AI, Part III open.substack.com/pub/alexan…
97
SOTA LLM architecture ablations for a novel encoder paper proceeding apace:
1
174
4. #Kimi Attention Residuals Block - github.com/MoonshotAI/Attent… - nuked my initial results, horrifically slow if using just the pseudocode (dropping it got me from 6-8 sec/it to 0.8~1.4 it/sec). Maybe needs better kernels (or a better coder than Opus, gemini, gpt and me).
1
134
5. Alt masking rates: jury out.
6. Most exciting <Censored> trick: The newest paper really does work much better, from being worse than MLM to outperforming it in a fair ablation.
Performant implementations are hard with no compute in medicine and biology 🤐
79
Dan Ofer (✈️ ICML-26 Was @Worldcon) retweeted
Apr 3
Every PI
11
103
812
155,853
I too have joined the autoresearch {trendy DL} thing crowd. Just hoping I can get a paper out on it sometime :)
100