the cost of shipping code went to zero
taste didn't
but "taste" sounds mystical and unfixable, so nobody teaches it. here's the unmystical version: taste is just an eval you haven't written down yet
how you choose what to measure is what matters
1/8
@charliermarsh Sometimes they're a good reason to work on something. When people say a market is "crowded," what that often means is that there's a real problem and none of the solutions are good enough yet.
@charliermarsh Sometimes they're a good reason to work on something. When people say a market is "crowded," what that often means is that there's a real problem and none of the solutions are good enough yet.
"Moramo da se izborimo za svoja građanska prava, sloboda nema cenu", rekao je na protestu na Slaviji profesor Hemijskog fakulteta, akademik Radomir Saičić. "Batinaju decu koja traže pravdu, a štite kriminalce koji aktivno pomažu ubicama", poručio je.
High-throughput platforms have found great application in process chemistry teams, where you need to debug reactions by exploring chemical space with a relatively standardized output to measure.
But at the scale of 96-384-1536 wells at a time, even the best chemists struggle to keep all results and hypotheses in mind, and have mostly converged on factorial-based designs.
Today, we are announcing Palladium, b12’s plate designer tool, which mimics the way chemists would think, hypothesize, test, analyze, and iterate on HTE conditions, and does so on scale.
What I find amazing is that you can go from SMILES strings to complex hypothesis testing in just a few clicks.
First results are very promising, stay tuned for experimental validation details soon.
In the quest toward automated synthesis, there's a massive gap between "this reaction looks good in paper" and actually running it.
Turns out, someone has to select the conditions: temperature, ligands, additives... the right conditions turn an impossible reaction into real product!
Today we’re announcing Palladium, our agentic solution for reaction condition and HTE plate design.
Instead of starting with blind optimization, Palladium leverages precedent experiments and substrate-specific reasoning to design a strong first plate.
Our beta users in academia and pharma have found conditions as high as 98%, and typically above 90% yield, on the first experiment, often with no follow-up experiments needed.
This is how at b12 Labs (YC S25) we accelerate real chemistry.
Read more https://t.co/xvSnnGcV6O
If you’re a chemist trying to accelerate difficult reactions, we’d love to talk.
📣 Announcing Terminal-Bench Science: benchmarking AI agents on real scientific workflows – now open for task contributions👇
https://t.co/MSPMwnbhVt
@AnthropicAI, @OpenAI, and @GoogleDeepMind use Terminal-Bench to evaluate AI on coding tasks. We're now extending it to scientific workflows.
1/6🧵
Great demos at Launch Live @ycombinator - remote control of robots, reinventing ultra sound, new transformer architectures for non-human tasks, and saving agents from dead-ends. YC events always feel like Im sent several years to the future.
The binders have bound! A few months ago, 9 human teams and 6 autonomous AI agents spent a single day designing protein binders against TREM2 on @muni_bio, a target implicated in Alzheimer’s Disease.
141 designs were submitted, 100 were synthesized and tested by @adaptyvbio, and 37 bound. And surprisingly, AI agents essentially matched human teams on hit rate.
These aren’t benchmark scores or simulated results, but real proteins designed in one day in SF and validated experimentally during the first large-scale test of muni, where teams ran 260 GPU jobs and generated a total of 4,176 binders.
We wrote about what we learned from the results, how well ipSAE worked as a scoring function, and how this hackathon reshaped what we’re building: https://t.co/uZswTolfIa
Great demos at Launch Live @ycombinator - remote control of robots, reinventing ultra sound, new transformer architectures for non-human tasks, and saving agents from dead-ends. YC events always feel like Im sent several years to the future.
@zhou_xian_ incredible work and progress! Would love to connect, at b12 labs we do chemical robotics control and would love to test your robots for some tasks in chemistry!!
Standout keynote this morning from @jrkelly from @Ginkgo , drawing a parallel between levels of flexibility and autonomy in cars and in labs.
The analogy: trains are highly automated but locked to a track. The model T was infinitely flexible but fully manual. Waymo is the breakthrough; flexible and autonomous. Labs are on the same trajectory. Traditional automation workcells are the “trains” of the lab; high throughput, low flexibility. The manual bench is the Model T. The frontier is the autonomous lab: as flexible as a scientist at the bench, as automated as a workcell.
The key takeaway: as we enter the programmable era of biology and chemistry, we’ll need more scientists, not fewer, to test hypotheses at a scale that was previously unthinkable.
@sama At https://t.co/FKNFkXaQO0 we built chemical reasoning engines and chemical robotics control agents that enable unknown small molecule synthesis in a wet lab - end to end automated. YCS25