@JocelynnPearl @EdisonSci Should use ours same capability and pay as you go
English
Devon Cayer
3.1K posts
Devon Cayer
@devcayer
entrepreneur, founder Om, founder of 1859, inc. — love intersection of AI and techbio — prefer to build.
San Diego, CA 参加日 Haziran 2009
2K フォロー中551 フォロワー
Welp, I guess this is the path... of LLMs? My favorite LLM for scientific research (FutureHouse's Platform) is now part of @EdisonSci and only offers a $200/month option or enterprise option. Nothing in between?
Sad because I was actually using it to help patients.
English
This is sad. I know as a politician these companies are going to spend a billion dollars against me for saying it but 🤷🏽♀️
Pervasive gambling is not good for society. It turns life into a casino, traps people in addiction & debt, surges domestic violence, and fosters manipulation.
Polymarket@Polymarket
We’re honored to announce MLB has named Polymarket as their Exclusive Prediction Market Exchange Partner. Polymarket 🤝 MLB
English
English
The sad part of building in stealth is when something phenomenal happens you can’t share it either.
Living through that moment.
English
@nimivashi15 You can share, but everyone will respond that you should open source it.
English
Devon Cayer がリツイート
We just released a big update to Rowan's Python API, with the goal of making it easier for (1) regular programmers (2) people using Claude Code/Codex, and (3) "AI scientists" to build atop Rowan.
Every workflow now returns a typed output result that mirrors the web interface:
English
It’s wild that you can go from hundreds of terabytes of data to maybe a gigabyte of high-performance models with the option to delete all prior data for max value capture
English
@devcayer I would guess someone is doing it! I could not find anything published though.
There is a nice opportunity to learn how to produce sequences that have multiple properties too...
English
There might be an interesting, tractable project to fine-tune plms on extremophiles from e.g., zenodo.org/records/171487…
There are thermostability plms, but there should also probably be pH, radiation, salinity, etc. I searched around a bit but didn't find this...
English
Would rather be funding investments in creation vs. destruction of countries.
open.substack.com/pub/elizabethg…
English
This reinforces my belief that bc the majority of data that will be generated on protein interactions will be sequence-based (ie no solved complexes), a major challenge and opportunity is building models that leverage seq data to outscale PDB-based models
Manifold Bio@ManifoldBio
@ManifoldBio and @NVIDIAHealth announce a joint study validating Proteina-Complexa, NVIDIA's latest BioNeMo model for protein binder design.
English
Any recommendations for an arbitrary de novo protein design tool that's ideally free? Every time I see de novo protein design it actually is antibody or binder design. I need it for things that have nothing to do with medicine. TYIA!
English
@Ronalfa Does that mean target idea then literature support to validate?
English
@Donatas_z @scrippsresearch Do one thing better than everyone else
English
@devcayer @scrippsresearch I was thinking about it . We still don’t know what “good” drug discovery is , so how AI will know ? 😉🥸
English
The ultimate test for AI - a thesis level advance, not a process optimization step after 25000 experiments
English