5 min read

Open-source AlphaFold rival drops

Open-source AlphaFold rival drops
Nº 01 · The Lede Hacker News Field report

Open-source model rivals AlphaFold

Open-source model rivals AlphaFold
Fig. IHacker News · Filed 28 May 2026.

Open-source structure predictor covers a billion proteins, matching AlphaFold-class accuracy with weights, training code, and inference pipeline all released, per Nature's writeup. The drop lands as the structure-prediction field has spent two years working around closed APIs and license terms. Coverage at billion-protein scale also includes microbial and metagenomic sequences that AlphaFold's public database underweights, expanding what counts as queryable structural biology.

Read the source

Critical flaw hits agent package
Fig. IIHacker News · Filed 28 May 2026.
Nº 02 Hacker News Agents · Infrastructure

Critical flaw hits agent package

Millions of AI agents are exposed by a critical vulnerability in a widely-used open-source agent package, per Ars Technica. The bug affects deployments across the agent ecosystem, including bio and clinical pipelines that pulled the dependency in without auditing it. Raises the floor for what agent platforms must ship: SBOM (software bill of materials) discipline and credential isolation become vendor criteria, not nice-to-haves.

Read more
RL agent learns chemistry rules
Fig. IIIarXiv · Filed 28 May 2026.
Nº 03 arXiv Agents · Infrastructure

RL agent learns chemistry rules

AtomComposer builds molecules atom-by-atom from first principles using reinforcement learning, with no training set of known compounds. The agent discovers valence and bonding rules on its own — anchors a counterargument to the assumption that chemical-space exploration requires a curated molecular corpus, and opens a path to scaffolds outside the drug-like distribution everyone trains on.

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

Agents reconstruct PK curves

A grey-box Transformer orchestrated by AI agents reconstructs sparse pharmacokinetic curves and initializes pharmacometric models from limited sampling. Narrows the gap between sparse-sample clinical PK studies and the dense-curve fitting that population modeling assumes.

Read
Nº 05 arXiv Field report

EEG foundation models get a yardstick

A multi-dimensional benchmark evaluates generalization in EEG foundation models across subjects, tasks, and recording setups. Becomes the new reference for whether neuro foundation models actually transfer — vendor claims of cross-dataset performance now have a checkable score.

Read
Nº 06 bioRxiv Structural biology · Protein design

Contrastive learning for protein fitness

SynFit predicts protein fitness across multiple objectives using synergistic contrastive learning, handling the tradeoffs between stability, binding, and expression in one model. Moves multi-objective protein design from sequential single-property optimization to joint training.

Read

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

Agentic Discovery  ·  Nº 25  ·  28 May 2026

Editor's Note

A billion proteins predicted, a million agents exposed, and a chemistry RL agent that skipped the textbook.

 

Nº 01 · The Lede  —  Hacker News  —  Field report

Open-source model rivals AlphaFold

Open-source model rivals AlphaFold

Fig. I  Hacker News · Filed 28 May 2026.

Open-source structure predictor covers a billion proteins, matching AlphaFold-class accuracy with weights, training code, and inference pipeline all released, per Nature's writeup. The drop lands as the structure-prediction field has spent two years working around closed APIs and license terms. Coverage at billion-protein scale also includes microbial and metagenomic sequences that AlphaFold's public database underweights, expanding what counts as queryable structural biology.

Read the source →

Why it matters

Resets the reference model for protein structure from a gated DeepMind asset to a fully open one — the entire downstream stack of docking, design, and annotation tools can now ship with structure prediction built in, not bolted on.

 

Nº 02  —  Hacker News  —  Agents · Infrastructure

Critical flaw hits agent package

Fig. II  Hacker News · Filed 28 May 2026.

Critical flaw hits agent package

Millions of AI agents are exposed by a critical vulnerability in a widely-used open-source agent package, per Ars Technica. The bug affects deployments across the agent ecosystem, including bio and clinical pipelines that pulled the dependency in without auditing it. Raises the floor for what agent platforms must ship: SBOM (software bill of materials) discipline and credential isolation become vendor criteria, not nice-to-haves.

Read more →

 

Nº 03  —  arXiv  —  Agents · Infrastructure

RL agent learns chemistry rules

Fig. III  arXiv · Filed 28 May 2026.

RL agent learns chemistry rules

AtomComposer builds molecules atom-by-atom from first principles using reinforcement learning, with no training set of known compounds. The agent discovers valence and bonding rules on its own — anchors a counterargument to the assumption that chemical-space exploration requires a curated molecular corpus, and opens a path to scaffolds outside the drug-like distribution everyone trains on.

Read more →

 

Also Filed  ·  Three Briefs from the queue

Nº 04  —  bioRxiv  —  Agents · Infrastructure

Agents reconstruct PK curves

A grey-box Transformer orchestrated by AI agents reconstructs sparse pharmacokinetic curves and initializes pharmacometric models from limited sampling. Narrows the gap between sparse-sample clinical PK studies and the dense-curve fitting that population modeling assumes.

Read →

Nº 05  —  arXiv  —  Field report

EEG foundation models get a yardstick

A multi-dimensional benchmark evaluates generalization in EEG foundation models across subjects, tasks, and recording setups. Becomes the new reference for whether neuro foundation models actually transfer — vendor claims of cross-dataset performance now have a checkable score.

Read →

Nº 06  —  bioRxiv  —  Structural biology · Protein design

Contrastive learning for protein fitness

SynFit predicts protein fitness across multiple objectives using synergistic contrastive learning, handling the tradeoffs between stability, binding, and expression in one model. Moves multi-objective protein design from sequential single-property optimization to joint training.

Read →

 

· · ·

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