←
Derek Alia929 posts
Derek Alia avatar

Derek Alia

@derekalia

Wizarding @muni_bio

◷ Joined December 2008▧ 98 Photos and videos
Derek Alia avatar
Derek Alia@derekalia·Oct 30

The factorioization of biology is happening. Expect faster/cheaper experimental validation and more novel drugs developed at a fraction of todays cost. My sense is that many people inside the drug development industry are further behind in understanding what's becoming possible than those on the outside looking in. Which leads to an interesting arbitrage, bound purely by ambition. Teams that embrace agentic workflows will soon look like an entirely different species of biotech, one that the old guard will struggle to comprehend. Some will interpret this negatively as the removal of human creativity from biological research, my view is opposite. I think this gives people with deep biological insight far more leverage. Freeing them from the tedious, repetitive work that consumes most of their context. Exciting times.

Kat@katyenkoOct 30

Binding is nice, but it’s getting boring. If you’re not thinking about biological function, you’re ngmi. At our last @muni_bio hackathon co-hosted with @onepot_ai, we set one ambitious goal: neutralize the latest threat of Ebola using small molecules. The wet lab data is in, and we’re excited to announce that teams were successful, including our agent that repurposed... an anti-itch gel?
 muni.bio/research/hits-from-…

Derek Alia avatar
Derek Alia@derekalia·6h

[2021] Strongly believe building new AI software engineering infra in SF is investor catnip. There are many problems with software development but the market for generating (quite sophisticated) code for cash is low on that list. Cycle through the ITT (“Infosys/TCS Test”): - Is the programming task something Infosys/TCS can do? - If so, is the AI equipped to deliver it at the quality, price and speed of I/T? This is very difficult. - If all of the above, can the AI handle new programming tasks with minimal additional training/time/expertise? Or are they accidentally building something with a fixed shape and the economics and pace of an outsourcing company? Think these differences in focus and capital allocation will show up clearly in growth over the next 6 months.

Strongly believe building new wet lab infra in SF is investor catnip. There are many problems with biotech but the market for generating (quite sophisticated) datasets for cash is low on that list. Cycle through the WPT (“WuXi/Pharmaron Test”): - Is the assay something WuXi/Pharmaron can run? - If so, is the team equipped to run it at the quality, price and speed of W/P? This is very difficult. - If all of the above, can the assay change shape with minimal additional build out/time/expertise? Or are they accidentally building something with a fixed shape and the economics and pace of a platform biotech? Think these differences in focus and capital allocation will show up clearly in growth over the next 6 months.

Derek Alia avatar
Derek Alia@derekalia·16h

This is awesome! Great work.

Claude was trained on protein sequences, at scale. We gave 20 language models the start of little-studied bacterial proteins and asked for the next amino acids. Only the latest Claude models predict them. Not GPT, not Gemini, not seven open-weight models.

Post attachment
Derek Alia avatar
Derek Alia@derekalia·Nov 16

👀👀👀👀

AI for science: autonomous design and synthesis of novel coolant materials.

1:57
Derek Alia avatar
Derek Alia@derekalia·Nov 11

Baller. If we can get a candidate, that’s AI pilled and builds, I’m sold.

I use C/C++ for all my projects. Debt Badge, Clucks Capacitor, homemade power-wall, sculpting software for my haptic interface. I suppose with AI assisting, the clunkiness of developing in C disappears and the efficiency shines. How soon until the language doesn’t matter?

Derek Alia avatar
Derek Alia@derekalia·Nov 4

new funding path unlocked. who needs vc when you can have the public foot the bill via taxes?!? this is crazy.

immad@immadNov 4

Something feels a bit iffy about California prop 38. Doing a popular vote proposed by a 3 person non-profit that gives said non profit $4.2B from the public! I hope Gary is great and knows how to spend $4.2B well because it’s done.

Post attachment
Derek Alia avatar
Derek Alia@derekalia·Nov 4

yes

So, could we take a small embedding model like ESM2-8M or ESMC-300M and ask Claude to identify the most relevant weights to modify/improve contact map prediction acc in order to mimic the behavior of larger models like ESMC-600M or ESM2-3B at a fraction of cost?

Emma avatar
⟳ Derek Alia retweeted
Emma@emmamariecryer·Nov 2

coolest email i have gotten since starting at @OpenRouter: "the compute you sponsored helped identify hits against Bundibugyo ebolavirus in our assays" you check out what @katyenko and her team were able to accomplish here:

Kat@katyenkoOct 30

Binding is nice, but it’s getting boring. If you’re not thinking about biological function, you’re ngmi. At our last @muni_bio hackathon co-hosted with @onepot_ai, we set one ambitious goal: neutralize the latest threat of Ebola using small molecules. The wet lab data is in, and we’re excited to announce that teams were successful, including our agent that repurposed... an anti-itch gel?
 muni.bio/research/hits-from-…

Derek Alia avatar
Derek Alia@derekalia·Nov 2

Lets go!!! I've been waiting for something like this.

We’re introducing Hydrogen-1.0, our first open-weight model for scientific research. We post-trained GLM-5.3 (@Zai_org) on our internal scientific data, lab records, and discovery trajectories. It scores 38.3% on FrontierScience-Research vs GPT-6-Astra’s 33.3% and scores 93.0% on GPQA-Diamonds. We believe science should be open, collaborative, and democratized, rather than proprietary. Only open weights let researchers inspect, adapt and build on the unprecedented speed of AI growth. There are many recent models named by chemical elements, but we believe Hydrogen-1.0 is the first aiming towards and open scientific superintelligence for discovery. 🧵

0:44
Derek Alia avatar
Derek Alia@derekalia·Oct 24

There is a moat, but it's no longer in software. It's the determination to succeed.

jack@jackOct 24

there is no moat

SPEC avatar
⟳ Derek Alia retweeted
SPEC@___4o____·Oct 22

There needs to be more discussion about the great replacement of Dungeons & Dragons/Warhammer 40K chads with pokercels and how it’s a microcosm of everything that went wrong in tech