Latent Space

Designing for Agents — Zeno Rocha, Resend

What Frontier Labs Buy: A Data Deep Dive with Handshake AI — Garrett Lord & Team

Deep Dive: Anthropic S-1 Myths and Mythos — Jamin Ball, Altimeter

Controlling Biology - Fabian Theis, Helmholtz Munich

⚡ Behind Latent Space: A Rare Look Inside the Pod | AIEi Shanghai

⚡️ The 2026 Software Factory Standard — Zach Lloyd, Warp.dev

Using Agents for Long-Horizon Financial Work — Aman Singh, Arca

⚡ Building ChatGPT for Finance — Joseph Kim, OpenAI's Head of Financial Services

Supabase’s Plan for a Billion AI-Built Apps

How Anthropic ships the #1 LLM in the World safely: Separating the Brain from the Hands

⚡ An Eval for AI Scientists That Knows the Answer Key - Ian Quigley, Leash Bio

Why AI Compute Comes in Rectangles — Erik Bernhardsson, Modal CEO

• Modal's CEO on reselling rigid GPU 'rectangles' as elastic compute →
• Modal CEO says a GPU price collapse worries him more than scarcity →
• Modal CEO on the sandbox product his team considered shutting down →
• 3 more articles from this episode are listed above.

The Recommendation Engine for Commodity Traders — Anekha Sokhal, Moshi

• Ex-energy trader builds Moshi, the market feed she wished she had →
• Moshi CEO says getting AI-built tools used is a recommendation problem →

AI for the Fortune 2000: How Tenex Builds an AI-Native SDLC — Alex Lieberman & CJ Hess

• Few big firms run agents outside engineering, Tenex's Lieberman says →
• Tenex's output-based pay produced a 4x gap, not 10x, says Lieberman →
• Tenex's Alex Lieberman doubts vertical AI apps and meets pushback →
• 3 more articles from this episode are listed above.

The Programmable Data Center — Evan Conrad, SF Compute

• Compute markets need physical delivery first, SF Compute CEO argues →
• Inside a GPU order book: SF Compute's CEO on how compute is priced →
• SF Compute CEO says a GPU market needs an operator that owns no GPUs →

Lab-in-the-Loop at Compute Speed — Ian Quigley, Leash Bio

• Leash Bio turns its binding data into a replayable lab for AI agents →
• Chemistry AI may learn who made a molecule, Leash Bio founder says →
• How Leash Bio tested AI binding prediction on unseen chemistry →
• 2 more articles from this episode are listed above.

Dyna-2: General Purpose Physical Agents with Hour-Long Autonomy — Jason Ma, Dyna Robotics

• How Dyna wraps a robot model in an agent for hour-long chores →
• Dyna's Jason Ma: Din Tai Fung napkins show robots beyond pilots →
• Dyna's Jason Ma on how robots learn without an internet of data →
• 3 more articles from this episode are listed above.

Why AlphaFold Didn't Solve Protein Folding — Pushmeet Kohli, Google DeepMind & Sal Candido, Biohub

• AlphaFold didn't solve protein folding, DeepMind's Kohli says →
• Biology AI needs the right data, not just more, two researchers argue →
• Kohli: biology AI must be calibrated, not necessarily interpretable →
• 1 more articles from this episode are listed above.

Synthesis Superintelligence: from Semiconductors to Superconductors — Periodic Labs’ Liam Fedus and Ekin Dogus Cubuk

• How Periodic Labs says it trains AI against a noisy physical lab →
• Periodic Labs bets failed experiments can teach AI to do science →
• Why Periodic Labs automates its lab one bottleneck at a time →
• 2 more articles from this episode are listed above.

Academia is for Ambition — Alex Zhang, MIT

• An MIT researcher guesses most agent-swarm search is wasted →
• Mainstream coding harnesses are all alike, Alex Zhang argues →
• Inside Prime Agent, an RLM harness where Python is the only tool →
• 2 more articles from this episode are listed above.