
fafo. ex: first engineer @metaformsai (evals, agents)
bangalore
Joined August 2013
- Tweets668
- Following861
- Followers759
- Likes50.2K
Siddish retweeted
I’m releasing OpenTTD-Bench, a benchmark for measuring frontier LLM agents on real-time, long-horizon tasks, based on a complex business simulation game OpenTTD, where models run competing transportation companies and must make the most amount of profit.
Siddish retweeted
new fav personal assistant use case unlocked!
i was annoyed that the roads near my house in HSR were dug up for MONTHS making it hard to even walk and it felt like there's nothing i could do about it!
early sept i asked faff to find a way to flag this to the authorities...
humans& and simile remind me of the davinci/InstructGPT days.
lots of fun experiments to run, so much to figure out before we get the chatgpt moment
For AI to work with us, it needs to understand us
Today, we're introducing Persimmon, the first large-scale model designed to realistically simulate how people talk and interact
1
10
915
Siddish retweeted
Super proud and excited to finally present LiteMol-1. This is our first foundation model from LiteFold, pre-trained from scratch.
Today, LiteMol-1 can generate small molecules, peptides, cyclic peptides, depsipeptides, peptides with ncAAs, macrocycles, and PROTACs.
Across our peptide and small-molecule evaluations, the model shows competitive results. In several settings, we are on-par with or better than frontier structure-based models, at a fraction of the generation cost.
But the part I find most interesting is that this is a model for agents. We have seen ourselves how much compute, and how many tokens it can take to generate good binders using frontier structure-based models. Sometimes you need to generate tens of thousands of designs just to get a handful worth taking forward.
Now put this inside an AutoResearch loop. The agent has to continuously parse structures, inspect PDB/CIF files, compare candidates, run evaluations, modify the design, and repeat the whole thing again. It becomes extremely expensive very quickly.
Sequence space gives us a very different interface. LLMs are much more efficient at inspecting and manipulating compact molecular representations like SMILES than repeatedly operating over full structural files.
So LiteMol-1 becomes something like an infinite molecular canvas for the agent. For a given target and objective, the model can continuously propose what a biomolecule could look like. The agent can inspect those generations, take inspiration from them, preserve certain regions, edit others, optimize them, score them, and generate again.
Generate → inspect → evaluate → edit → generate again.
There is another problem I care a lot about. Most molecule design models today are heavily optimized around binding. But binding is only one part of whether something eventually becomes a therapeutic. What about ADME? Toxicity? Selectivity? Solubility? Membrane permeability? Synthesizability?
For this, we also built a Monte Carlo Tree Search-based multi-objective generation framework around LiteMol-1. Instead of combining everything into one score, the search keeps multiple strong candidates, each balancing the desired properties in a different way.
As our scoring functions and verifiers get better, the generation system gets better too. We can start steering molecules not just toward “binds well”, but toward a broader therapeutic design specification.
The bottleneck slowly moves from simply generating molecules to having sufficiently good verifiers and scoring functions to tell us what is actually worth generating. Check out our technical research blog post for all the details.
At LiteFold, our research is focused on engineering biomolecules and building systems that can carefully forecast their pre-clinical success.
To stay updated on our research, follow LiteFold.
Cheers!
Today, we are incredibly excited to present LiteMol-1, our very first foundation model from LiteFold.
Structure-based models like BoltzGen, O-Design, and RFdiffusion have become the de facto standard for designing biomolecules. However, they are expensive to run at scale. In many campaigns, we have to generate tens of thousands of designs and rigorously filter them down to a handful of top candidates.
More importantly, most molecule design systems are primarily optimized around binding. But what about everything else that makes a molecule a useful therapeutic: ADME, toxicity, drug-likeness, membrane permeability, synthesizability, selectivity, and more?
That’s where we introduce LiteMol-1, our first Multi-Molecule Foundation Diffusion Language Model, pre-trained from scratch. With a single set of weights, LiteMol-1 can conditionally generate small molecules, peptides, cyclic peptides, depsipeptides, peptides with ncAAs, macrocycles, and PROTACs.
You can generate molecules unconditionally or condition generation on a protein target sequence. You can in-paint molecules, preserve parts of an existing scaffold, generate non-canonical peptides, cyclize designs, and continuously edit and regenerate them.
We also introduce a Monte Carlo Tree Search framework for multi-objective molecular design. Instead of trying to make one molecule perfect at everything, the search keeps a set of promising molecules, each with different trade-offs across properties. This matters because molecular design is fundamentally not a single-objective optimization problem.
Finally, and probably the part we are most excited about: this is a model for Agents. LLMs are not particularly efficient interfaces for repeatedly reasoning over thousands of large PDB/CIF files inside an AutoResearch loop. Sequence space is different. SMILES and molecular sequences are compact, editable, and much easier for an agent to inspect, compare, modify, and reason over.
So we treat LiteMol-1 as an infinite molecular canvas.
The model generates possibilities. The agent takes inspiration from them, evaluates them, edits them, optimizes them, and generates again. The AutoResearch loop continues until it finds candidates that satisfy the given design objectives; while bringing in expensive structure prediction, docking, or simulation only when they are actually needed.
Our evaluations across peptide and small-molecule generation show that LiteMol-1 is competitive with, and in several settings on-par with or better than, frontier structure-based and sequence-based models, while operating at a fraction of the generation cost and time.
And this is only the first step. Check out our technical research blog post in the comments to learn more about LiteMol-1.
Siddish retweeted
Looks like this blew up, we've been able to help over 40k people in the last 24 hours.
We're funding this out of our pocket and we've spent >$500 in the last two days which is becoming unsustainable :((
Would REALLY appreciate any offers to help with the exploding costs here, my DMs are open.
(Also accepting OpenRouter API keys!)
The ECI is doing a nationwide cleanup of the rolls but the raw data has TONS of issues, from bad translation to typos.
@NotNotSaumya and I built a search to fix this. Provide as much detail as you can and we filter through thousands to find you!
Examples from folks using it:
Siddish retweeted
If you're NOT in Bangalore and working on the public good on any of our favorite problem areas, come to Bangerlore
We at @_lagrangepoint will help with flights and accommodation, and help you meet fellow practitioners. Just DM me!
- urban quality of life / civic tech
- food toxicity
- deep tech talent
If you know someone that fits, send them my way, and I will gift you some really nice socks
Siddish retweeted
⏰ 3 days left until the deadline for the Conference for AI Scientists (CAISc) 2026.
With awards incl. model and compute credits over $20,000, this is a great chance to get feedback on ongoing research.
Our ask - use an LLM or AI system in some part of your research workflow.
Siddish retweeted
built this in the last 4 weeks or so. it's a full fledged market research delivery ops simulator with operations, finance, demand-gen, and HR modeled in. wanted a good way to teach operational concepts + simulate what agents do to an operational chain
mrxsimulator.com/
Siddish retweeted
LiteFold just got featured in @YourStoryCo
It’s been 8 months since Cory and I officially started @try_litefold.
We began by building better interfaces and workflows for computational biology, followed by Rosalind, our AI Co-Scientist for therapeutic design.
Now we’re gradually moving toward developing our own in-house scientific and AI research focused on next-generation therapeutic models and research environments and partnerships.
Still very early. Really cant wait to share whats coming.
Story link in the comments
Siddish retweeted
Excited to release the Ultimate guide to RL environments!
Definitions of RL environments differ wildly in the LLM era, so we spent the last month building several RL environments across 6 different frameworks, domains and complexities to map out which are easiest to build with and which can be scaled to 1000s.
Siddish retweeted
Introducing RL Commons.
An open research initiative for the reinforcement learning era.
Shared compute. Open benchmarks. Collaborative work.
Founding phase Project Aster is now open.
Siddish retweeted
Frontier LLMs can do a lot—but can they write good flashcards?
Turns out: not yet! @andy_matuschak and I created an eval for flashcard generation and found surprisingly poor results.
Worse, newer models aren’t helping: GPT 5.4 performs worse than 5.2, Opus 4.7 worse than 4.6.
Siddish retweeted
this is our thesis. it's about squid, MRI machines, and why we think math is what's holding science back.
principialabs.org/blog/the-b…






























