fueled by randomness ml research @cusp_ai, ex-{@MSFTResearch,@GoogleDeepMind,@MPI_IS} born @ 353 ppm, he/him, block toxicity

Berlin
Joined April 2019
Super proud to see this finally out! Amazing work with @n_gao96 and my other lovely colleagues from @cusp_ai ! Looking forward to seeing what the community will actually build with this! ☕️☕️☕️🚀
Molecular simulation is the backbone of drug discovery and materials science, but it’s notoriously complex. Today, we’re excited to open source kUPS - a molecular simulation engine built for the AI era, optimized for GPU in collaboration with @nvidia.
1
2
22
2,159
Jonas Köhler retweeted
oh Gott Leute, ich habe gerade den Wikipedia-Artikel über Deutschland gelesen und da steht die Linken waren gar nicht Schuld an Hitler????
84
79
1
2,439
26,103
Jonas Köhler retweeted
Wiederhole mich: Man kann die Linke politisch beknackt finden, aber für Demokratie & Verfassung ist sie keine Gefahr. Im europäischen Vergleich ist die deutsche Linke sogar moderat, das gilt auch für Haltung zu Israel. AfD hingegen ist eine der rechtsextremsten Parteien Europas.
2,827
725
159
7,155
306,649
Jonas Köhler retweeted
ich bin gerade in berlin und sie stellen schon die erste lenin statue auf
29
333
23
8,025
105,128
Jonas Köhler retweeted
Most people without experience fail to appreciate just how hard (i.e., nearly impossible) it is to simulate the physical world with high fidelity. It's exactly the same as people who've never been in a wet-lab assuming that you can just think your way to cooking up a new virus.
1
9
70
3,390
Jonas Köhler retweeted
Interested in pursuing a PhD with me at @UvA_Amsterdam on mechanistic interpretability, AI for science, or AI safety? NWO Mozaiek is funding PhD students with non-Western migration backgrounds: nwo.nl/en/calls/mosaic-30
4
11
53
3,227
Jonas Köhler retweeted
it's not "out-of-control" at all. openai had all the control in the world (i.e. they could always literally turn off all their machines across all data centers.) they decided not to based on carefully calculated trade offs between finance-economy-reputation-customers-etc. just like any other co'a e.g. anthropic
29
102
16
858
71,148
Jonas Köhler retweeted
Yes let’s all listen to p(doom) predictions from 25yo boys whose frontal lobes formed yesterday
1
1
7
562
Jonas Köhler retweeted
Navier Stokes represents a new moment for AI. Hopefully, OpenAI will turn its attention to problems of societal importance, like DeepMind has, and to the mathematical foundations of AI alignment, which are sorely lacking.
With AlphaFold we mapped the protein universe - now with AlphaGenome Atlas we’re charting the human genome. It can predict the impact of all 9 billion possible single-letter DNA variants, helping scientists better understand disease. Freely available for academic research: alphagenome.google/atlas
7
1
108
12,223
Seems like ala2 is not considered a threat to humanity anymore with Fable 5.1 :)
The hypocrisy of @AnthropicAI @claudeai rolling out it's own bio subsidiary while blocking work on basic FOSS MD tools and established toy systems like alanine dipeptide (for safety!) is ridiculous in the best reading and outright evil in the worst. AI for good - yeah right.
1
7
1,044
Jonas Köhler retweeted
Celebrating my co-founder Chad Edwards making the #TIME100AI list today! Chad is a once in a generation CEO: tenacious, innovative and kind. Building [tag]CuspAI alongside him has been an incredible journey. Read Billy Perrigo’s profile in @TIME: time.com/collection/time100-…
3
5
67
7,415
Jonas Köhler retweeted
CuspAI co-founder & CEO Chad Edwards has been named on this year’s TIME100 AI list. Discovering materials to power clean energy & next-gen compute takes bold leadership, congrats @ac_edwards_1! If these hard challenges sound like your thing, we’re hiring… #TIME100AI
1
1
18
1,221
Jonas Köhler retweeted
Projects promoting programming in "natural language" are intrinsically doomed to fail. - Edsger W.Dijkstra, June 18th 1975
127
289
47
5,681
199,863
Certain lab given recent upcoming release rumors.
they’re calling it the wettest video ever made
3
855
Jonas Köhler retweeted
This is exactly how AI hype turns real science into fake news. Yes, Moderna used AI and computational algorithms in this cancer program, especially to analyze tumor mutations and select neoantigens. But this is NOT a story about ChatGPT, LLMs or the recent generative AI boom suddenly curing cancer. Moderna and Merck started this program in 2016. The first patient was dosed in 2017. The technology has been developed and clinically tested for years. The actual breakthrough is personalized mRNA cancer therapy + Keytruda showing positive Phase 3 results in melanoma. That is already huge. There is no need to rewrite it as “AI cured cancer.” I am very pro AI, but pretending every scientific breakthrough happened because of today's AI hype makes the whole field look less credible. In an era where anyone can generate convincing nonsense in seconds, accurate and authentic posts will become much more valuable.
AI labs: give us billions, we’ll cure cancer Moderna:
18
101
2
613
43,997
Claude is just a science wrapper 🍿 :)
5
375
Jonas Köhler retweeted
New work I'm very excited about: Physics of Agents!🤖 As AI agents become more prevalent, they'll interact and influence one another. Can we predict the collective behavior that emerges? Surprisingly, their dynamics follow compact, predictive laws of statistical physics🧵
47
194
32
1,330
158,353
Jonas Köhler retweeted
a nice demonstration of Claude Science, but worth clarifying that the design is not "done by Claude" but by orchestrating tool calls of open-source, task-specific protein design models: PXDesign, RFdiffusion, Genie, BoltzGen, etc I think the direction of LLMs using biology-specific models is a good one!
Many drugs work by binding to a specific target in the body and blocking or changing what it does. An important first step in the drug development process is designing a molecule that can bind tightly to its target. Traditionally, that's meant weeks or months of expert work per target, sifting through a large number of candidates to identify the few that work. We wanted to test if Claude could successfully design novel protein binders from scratch (also called de novo design). With a protein design prompt written by a human expert, Claude autonomously designed protein binders against 14 out of 15 targets. We then worked with Adaptyv Bio and Twist Bioscience, who independently built and tested the proteins Claude designed.
31
84
12
759
76,493