Sebastian Schmon

Saying there is a "distribution shift" between train and test data is a great marketing ploy in #machinelearning to make overfitting somehow the data's fault, not the model's or modellers'. 🙃

Love seeing external teams succeed—Onava just published their Latent-X1 binder results! Using the Latent Labs Platform to design proteins in the web browser, the team around Sharrol Bachas was able to design binders against a particularly challenging immune target. They completed design-lab loops in just 7 days! Key highlight: Nanomolar-strength binders with high target specificity, outperforming that of other approaches in their lab comparison. Of course we just launched Latent-X2 now enabling drug-like antibody design. Reach out if you'd like to explore what that could mean for your pipeline. See their preprint here: biorxiv.org/content/10.648…

Super proud of the team's achievement 🚀 Shipping our second model Latent-X2 that allows users to design antibody binders on our platform. Check the below thread for more details 👇

5 months from Latent-X1 to Latent-X2. AI-generated antibodies with drug-like developability and low immunogenicity in human panels — zero-shot. Available now for selected partners.

Amazing work by my friends at latent labs!

Latent Labs Announces Latent-X2: AI-Generated Antibodies With Drug-Like Developability and Low Ex Vivo Immunogenicity sfbn.org/san-francisco-… #news #biotech

New latent labs report on de novo antibody design closely follows what we’ve seen from Chai and Nabla. Promising results, and I respect that they admit that half of the targets were unsuccessful with their approach

Introducing Latent-X2 — AI-generated antibodies with drug-like developability and low immunogenicity in human panels, zero-shot. Technical report: tinyurl.com/latent-X2-repo… Blog: latentlabs.com/latent-x2 Apply for access: partnerships@latentlabs.com

Another packed room and great company, thanks everyone for coming to our second In Silico event. Special thanks to @amyxlu @SimMat20 @SeBayesian and @DominiqueLaLo for the great talks! And our sponsors and hosts @RenPhil21 and @IDEALondon Look out for the next edition soon 👀

Proud to share that our founder Simon Kohl was named in the @observer's 2025 A.I. Power Index. @saakohl's recognition reflects both his groundbreaking contributions and enduring influence in AI's transformation of biology. See the full roster of A.I. leaders: observer.com/list/2025-ai-p… #ObserverPowerIndex

In Silico is back for round #002. Come see great talks on the future of protein and molecular design from @amyxlu @SimMat20 @SeBayesian and @DominiqueLaLo. October 1st 6pm at @IDEALondon luma.com/txswapqo

Showcasing Latent-X's breakthrough macrocycle designs. Our AI achieved an unprecedented 91-100% hit rate with single-digit micromolar binding affinities—functional binders on every single target tested. Explore our binders without logging in: platform.latentlabs.com

It's been a fantastic first few days of public usage of the Latent Platform! The team has been providing white glove user support - for feedback or requests, you can reach us through our feedback channels: form.typeform.com/to/P4vxcyT2?ty… Also, check out our sharing feature below: You can share design batches for others to view right in the platform, without even needing to sign in!

🧬 Latent-X is our first frontier model for protein design. For mini-binders, our AI model achieved picomolar binding affinities — the strongest reported vs prior methods in head-to-head lab validation. Explore our best binders here (no login required): platform.latentlabs.com More details on the breakthrough in this🧵

I'm really excited about what we build here, the machine learning, the science, but also the platform! Proud to be part of an amazing team!

Introducing Latent-X — our all-atom frontier AI model for protein binder design. State-of-the-art lab performance, widely accessible via the Latent Labs Platform. Free tier: platform.latentlabs.com Blog: latentlabs.com/latent-x/ Technical report: tinyurl.com/latent-X

@Latent_Labs comes out of stealth today with $50M funding. Our goal? To push the frontiers of generative biology, giving partners instant access to tools capable of accelerating drug design. Every biotech or pharma company searching for the best therapeutic molecules understands the role AI can play - but not all are in a position to develop their own advanced models. That’s where @Latent_Labs comes in.

@hayou_soufiane Exactly, and fine-tuning examples are crafted with question and answer, not with "I don't know". Similarly, there is a reason why con-artists exists and why we elect politicians who lie and or promise the impossible.

@SeBayesian It's just a reflection of Human behavior (you hardly find "i don't know" on forums)

Or: "Why are Chatbots trained to tell us what we want to hear so reluctant to say 'I don't know'?" It's a mystery. 🤔

@arankomatsuzaki It's so over. And at this rate it's gonna be overer pretty soon.

The leap from o1 to o3 is exponential, completely bypassing o2. If this pattern holds, o3 won’t lead to o4—it’ll jump straight to o9.

If you believe (=0) do you also believe (sinx)/n = six? In maths you can't just leave out the definitions and start with the proof. Do you mean both x to be exactly the same entity or do you mean one fixed but unknown quantity for every letter in the alphabet, i.e. (x'-x)?