Genesis Molecular AI: Diffusion models for drug discovery, with Llama lead Sergey Edunov and PEARL's zero-shot OpenBind win
Latent Space interviews Evan Feinberg and Sergey Edunov (formerly Meta's Llama lead) about Genesis Molecular AI, a startup applying diffusion models to drug discovery. The conversation covers PEARL's zero-shot performance on the OpenBind benchmark and the broader implications of co-folding models crossing accuracy thresholds for molecular design. The piece argues that the most interesting diffusion research is happening in scientific domains rather than language modeling.
Related guides (3)
Related events (8)
PLAID: Repurposing Protein Folding Models for Multimodal Protein Generation with Latent Diffusion
PLAID is a generative model that simultaneously produces protein 1D sequences and 3D all-atom structures by learning a diffusion model over the latent space of ESMFold, a protein folding model. It requires only sequence data for training—leveraging databases 2-4 orders of magnitude larger than structure databases—and decodes structure at inference via frozen folding model weights. The approach supports compositional prompting by function and organism, addressing practical drug-design constraints like humanization and solubility. A companion compression model, CHEAP, addresses the high-dimensionality of transformer latent spaces to make the diffusion training tractable.
Simon Willison on DiffusionGemma
Simon Willison covers DiffusionGemma, a diffusion-based language model in the Gemma family from Google. The post appears to be commentary or a brief note on the model's release or capabilities. Diffusion-based LLMs represent an active area of research as an alternative to autoregressive generation.
ESMFold2: The Bitter Lesson is Coming for Proteins — Alex Rives, BioHub
A Latent Space interview/commentary piece featuring Alex Rives of BioHub discussing ESMFold2 and the application of the 'bitter lesson' (scale and general methods beating hand-crafted inductive bias) to protein structure prediction and biology. The piece covers the tension between dataset scale versus domain-specific inductive bias in biological ML, and touches on world models and programmable biology. This represents a significant perspective from a leading researcher in protein language models on the next generation of biological foundation models.
Interpretability study of DiffusionGemma reveals novel diffusion-specific reasoning phenomena
Researchers investigate the reasoning transparency of DiffusionGemma, a diffusion-based language model, decomposing transparency into variable and algorithmic components. They show that mapping information through an interpretable token bottleneck reduces DiffusionGemma's opaque serial depth from 28.6X to just 1.1X that of autoregressive Gemma 4, with no performance loss. Interpretability case studies uncover diffusion-specific phenomena including non-chronological reasoning, token smearing, and intermediate-context reasoning. Monitorability tests find DiffusionGemma comparable to Gemma 4, suggesting diffusion LMs are not inherently less amenable to safety oversight.
GLM-5.2 claims top frontend coding performance; IndexShare speculative decoding introduced
A Latent Space AI news digest highlights GLM-5.2 as a new open-weights model claiming top performance on frontend coding tasks. The digest also covers IndexShare, a technique for speculative decoding. The body is truncated but the headline signals a notable open-weights model release and an inference optimization development.
DeepMind announces DiffusionGemma with 4x faster text generation
DeepMind published a blog post introducing DiffusionGemma, a diffusion-based variant of the Gemma model family claiming 4x faster text generation. The announcement suggests a departure from standard autoregressive decoding in favor of diffusion-based generation. If the claims hold, this could represent a meaningful inference efficiency advance for the Gemma line.
The Batch: Jalapeño inference chip, Fugu multi-agent system, Claude Tag, Robin bio-agent, and Getty-OpenAI deal
OpenAI and Broadcom announced Jalapeño, OpenAI's first custom inference chip, designed in nine months with AI-assisted design and showing better performance-per-watt than current accelerators; engineering samples are already running GPT-5.3-Codex-Spark with datacenter deployment planned by end of 2026. Sakana AI released Fugu, a multi-agent routing system that scored 73.7% on SWE-Bench Pro, outperforming Claude Opus 4.8 and GPT-5.5 while remaining below the inaccessible Fable 5. Additional items cover Anthropic's Claude Tag Slack integration for async team collaboration, Seedance 2.5 video model improvements, the Robin autonomous biology research agent that identified a novel drug candidate, and a Getty Images licensing partnership with OpenAI.
AlphaGenome: DeepMind's Unified DNA Sequence Model for Regulatory Variant-Effect Prediction
DeepMind has introduced AlphaGenome, a new unified DNA sequence model designed to advance regulatory variant-effect prediction and improve understanding of genome function. The model is now available via API, making it accessible to researchers. AlphaGenome represents a significant step in applying large-scale AI to genomics, particularly for interpreting non-coding regulatory regions of the genome.