6 min read

Big pharma picks its AI model vendor

Big pharma picks its AI model vendor
Nº 01 · The Lede X Field report

BMS bets on Chai's models

BMS bets on Chai's models
Fig. IX · Filed 21 Aug 2026.

BMS is deploying Chai's models across therapeutic discovery, cofounder Joshua Meier said, one of the clearest signals yet that AI structure and interaction models have moved from evaluation to production inside big pharma. Chai Discovery builds models that predict biomolecular structure and binding, the layer sitting between target selection and a real candidate molecule. The announcement's framing is blunt: frontier drug programs require frontier capabilities. What shifts is the buying question — pharma is no longer asking whether these models clear a benchmark, but whose models get written into the discovery pipeline. Model access, not model existence, becomes the competitive variable in early discovery.

Read the source

Anthropic talks binder design
Fig. IIX · Filed 21 Aug 2026.
Nº 02 X Field report

Anthropic talks binder design

Anthropic walked through binder design in a public post explaining that many drugs work by binding a target tightly, and that designing that first tight-binding molecule is the hard opening step of a program. The content is introductory; the sender is the news. A general-purpose AI lab writing directly to a drug-discovery audience rather than a developer one signals molecular design being treated as an application area in its own right, which changes who biology gets to count as a discovery-tools supplier.

Read more
Agent evolves its own algorithms
Fig. IIIbioRxiv · Filed 21 Aug 2026.
Nº 03 bioRxiv Agents · Infrastructure

Agent evolves its own algorithms

Autonomous algorithm evolution arrives in drug development in a new bioRxiv preprint, describing a system that generates and improves its own algorithms rather than executing fixed ones. Detail beyond the design claim is thin, but the direction is the notable part: most discovery tooling treats the method as fixed and the data as the variable. Handing method development to the system itself shifts the capability question from how good today's algorithm is to how fast the algorithm can be improved.

Read more
Also Filed · Four Briefs from the queue
Nº 04 bioRxiv Agents · Infrastructure

Cryo-EM agent learns to stop

CryoForge knows when to stop building. The self-correcting agent handles cryo-EM model building and decides when further refinement helps versus when to halt, putting an explicit stopping criterion into a workflow where iteration still runs largely on human judgment.

Read
Nº 05 arXiv Field report

Feasibility built into peptide design

FAR-DPO steers peptide generators toward molecules that can actually be made, folding feasibility into direct preference optimization, the training method that pushes a model toward preferred outputs. Synthesizability moves inside the objective for cyclic-peptide design rather than acting as a downstream filter.

Read
Nº 06 arXiv Field report

Test-time tuning for virtual screening

PETA adapts screening models at test time by updating only a small share of parameters, aimed at the distribution shift that degrades virtual screening on unfamiliar target classes. Cheap adaptation lowers the retraining cost that keeps screening models pinned to the chemistry they were trained on.

Read
Nº 07 OpenAI Field report

OpenAI's GPT-5.6 builder guide

OpenAI published a builder's guide to GPT-5.6, covering model selection and new Responses API capabilities aimed at cheaper agents. Cost per agent run is the binding constraint on long-horizon discovery loops, so routing guidance lands on the budget line.

Read

Reply with your discoveries. A human reads them. Forward freely.

Agentic Discovery  ·  Nº 82  ·  21 Aug 2026

Editor's Note

BMS names a structure-model supplier, Anthropic starts writing to chemists, and an agent learns when to quit.

 

Nº 01 · The Lede  —  X  —  Field report

BMS bets on Chai's models

BMS bets on Chai's models

Fig. I  X · Filed 21 Aug 2026.

BMS is deploying Chai's models across therapeutic discovery, cofounder Joshua Meier said, one of the clearest signals yet that AI structure and interaction models have moved from evaluation to production inside big pharma. Chai Discovery builds models that predict biomolecular structure and binding, the layer sitting between target selection and a real candidate molecule. The announcement's framing is blunt: frontier drug programs require frontier capabilities. What shifts is the buying question — pharma is no longer asking whether these models clear a benchmark, but whose models get written into the discovery pipeline. Model access, not model existence, becomes the competitive variable in early discovery.

Read the source →

Why it matters

Structure and binding models just became a procurement line inside a top-tier pharma, which resets the bar for every discovery-AI vendor from "publish a benchmark" to "name a partner running programs on it."

 

Nº 02  —  X  —  Field report

Anthropic talks binder design

Fig. II  X · Filed 21 Aug 2026.

Anthropic talks binder design

Anthropic walked through binder design in a public post explaining that many drugs work by binding a target tightly, and that designing that first tight-binding molecule is the hard opening step of a program. The content is introductory; the sender is the news. A general-purpose AI lab writing directly to a drug-discovery audience rather than a developer one signals molecular design being treated as an application area in its own right, which changes who biology gets to count as a discovery-tools supplier.

Read more →

 

Nº 03  —  bioRxiv  —  Agents · Infrastructure

Agent evolves its own algorithms

Fig. III  bioRxiv · Filed 21 Aug 2026.

Agent evolves its own algorithms

Autonomous algorithm evolution arrives in drug development in a new bioRxiv preprint, describing a system that generates and improves its own algorithms rather than executing fixed ones. Detail beyond the design claim is thin, but the direction is the notable part: most discovery tooling treats the method as fixed and the data as the variable. Handing method development to the system itself shifts the capability question from how good today's algorithm is to how fast the algorithm can be improved.

Read more →

 

Also Filed  ·  Four Briefs from the queue

Nº 04  —  bioRxiv  —  Agents · Infrastructure

Cryo-EM agent learns to stop

CryoForge knows when to stop building. The self-correcting agent handles cryo-EM model building and decides when further refinement helps versus when to halt, putting an explicit stopping criterion into a workflow where iteration still runs largely on human judgment.

Read →

Nº 05  —  arXiv  —  Field report

Feasibility built into peptide design

FAR-DPO steers peptide generators toward molecules that can actually be made, folding feasibility into direct preference optimization, the training method that pushes a model toward preferred outputs. Synthesizability moves inside the objective for cyclic-peptide design rather than acting as a downstream filter.

Read →

Nº 06  —  arXiv  —  Field report

Test-time tuning for virtual screening

PETA adapts screening models at test time by updating only a small share of parameters, aimed at the distribution shift that degrades virtual screening on unfamiliar target classes. Cheap adaptation lowers the retraining cost that keeps screening models pinned to the chemistry they were trained on.

Read →

Nº 07  —  OpenAI  —  Field report

OpenAI's GPT-5.6 builder guide

OpenAI published a builder's guide to GPT-5.6, covering model selection and new Responses API capabilities aimed at cheaper agents. Cost per agent run is the binding constraint on long-horizon discovery loops, so routing guidance lands on the budget line.

Read →

 

· · ·

Reply with your discoveries. A human reads them. Forward freely.