Issue No. 37Week ending Sunday, September 13, 2026434 episodes · 1825 articles
The Throughline ↓
The Podcast Summary.

40 hours of podcasts, in 5 minutes.

Guest

R.J. Honiki

R.J. Honiki appears in 6 full episodes we cover on Latent Space. Below is what each conversation covered, with a key takeaway per article. Every quote in the articles is verbatim and timestamped to the source video.

6 episodescovered
23 articleswith timestamped quotes
AILatent Space

🔬 Why Transformers Hit a Wall the Moment Physics Shows Up — Anima Anandkumar, Caltech

Caltech professor and former NVIDIA AI research leader Anima Anandkumar discusses why traditional deep learning and transformers fail when applied to complex physical systems due to extreme resolution requirements and s…

  • FourCastNet delivers weather forecasts comparable to supercomputer numerical models while running tens of thousands of times faster on a single GPU. Read →
  • Standard transformers hit a computational wall on high-resolution physics simulations because self-attention scales with quadratic complexity across dense 3D and 4D grids. Read →
  • Caltech researcher Anima Anandkumar achieved a 1,000,000x speedup modeling tokamak plasma evolution compared to traditional physics simulations. Read →
AILatent Space

🔬Biology Is Turning Into Software — Matt McPartlon and Neil Patil, Chai Discovery

This episode features Matt McPartlon and Neil Patil from Chai Discovery, a protein design startup, discussing their AI models (Chai 1, 2, and 3) for protein and antibody design. They elaborate on their unique partnershi…

  • Traditional drug discovery follows a "waterfall model" where each stage, from target to optimization, takes months or years and costs a fortune to even try new things, according t… Read →
  • Neil Patil of Chai Discovery, a protein design startup, identifies "talent obscurity" as the biggest bottleneck for AI in biology, noting top ML talent often gravitates to LLMs or… Read →
  • Chai Discovery acts as a "neutral software factory" for medicines, focusing solely on providing AI models and a product platform rather than developing its own drugs. Read →
AILatent Space

🔬Causal Models Need Causal Data - Xaira’s X-Cell model (Bo Wang & Ci Chu)

This episode delves into Zera Therapeutics' AI drug discovery platform, focusing on their X-Cell model. Guests Bo Wang and Ci Chu explain how X-Cell leverages high-throughput causal data generation (Perturb-seq) and a n…

  • Academic freedom isn't a luxury, it's a speed advantage. Bo Wang argues university labs, free from immediate profit pressures, are the wellspring for radically new concepts, like… Read →
  • Even with advanced AI drug discovery platforms like Zera Therapeutics’ X-Cell model, the current bottleneck for virtual cell modeling isn't just compute power or algorithms, but t… Read →
  • Zera Therapeutics' X-Cell model uses a diffusion language model architecture to predict cellular responses, departing from traditional auto-regressive (GPT-like) methods. Read →
AILatent Space

🔬 "The Most Innovative Diffusion Research Is Happening in Drug Discovery, Not Image Generation"

This episode explores how Genesis Molecular AI is leveraging advanced AI, particularly diffusion models, to revolutionize drug discovery by achieving unprecedented accuracy in protein-ligand structure prediction. Guests…

  • The highest leverage for AI in healthcare isn't identifying new disease targets or optimizing clinical trials; it's in the drug discovery and design process itself. Read →
  • The 1 Angstrom Cliff: Genesis Molecular AI's Pearl model demonstrates that for drug discovery, 1 angstrom (Ã…) resolution isn't just nice-to-have; it's the absolute minimum thresho… Read →
AILatent Space

Scaling Past Informal AI - Carina Hong, Axiom Math

This episode features Carina Hong, CEO of Axiom Math, discussing her company's vision for formal verification as the foundation for superintelligence and AGI, backed by a significant Series A funding round. Hong details…

  • AI, particularly Lean-based systems, struggles with highly creative mathematical domains like combinatorics because the necessary steps are often too intuitive and "quite creative… Read →
  • Axiom Math secured a $200 million Series A funding round, valuing the company at $1.6 billion, to advance "verified AI." Read →
  • Axiom Math, fresh off a $200 million Series A, aims to move formal verification beyond mere error correction, focusing instead on scaling "brilliance" to achieve superhuman math p… Read →
AILatent Space

🔬 The Bitter Lesson is Coming for Proteins - Alex Rives, BioHub

This episode features Alex Rives from Biohub discussing the application of the 'bitter lesson' (scaling laws and empirical evidence) to protein biology through the ESMC language model. Rives explains how massive dataset…

  • The "Bitter Lesson" applies to biology: Alex Rives, a key figure at Biohub, champions the principle that empirical scaling with massive data often outperforms human intuition or h… Read →
  • ESMC's Blind Insight: The ESMC protein language model, trained only to predict amino acids from sequences, developed a "world model" of proteins so accurate it intrinsically under… Read →
The Sunday Email

Get next Sunday's issue in your inbox.

40 hours of podcasts, distilled into one 5-minute read. Free, every Sunday morning.

One email a week. Unsubscribe with one click.