Latent Space: The AI Engineer Podcast
Latent.Space
8 episodes
[One Million AI Engineer Special] How Stanford Teaches AI-Native Engineering — Mihail Eric, Monaco
8 Nov · 2 h 1 min
- Mihail Eric's Stanford CS146S returns fall 2026, and week one has students build a coding agent from scratch in about 200 lines.
- Eric says code review is now a bottleneck, with even software factories claiming only around 30 percent of contributions are fully agent-driven.
Deep Dive: Anthropic S-1 Myths and Mythos — Jamin Ball, Altimeter
7 Nov · 1 h 42 min
- Jamin Ball expects Anthropic to begin public trading between November 15 and 20.
- Anthropic's S-1 shows 85% gross margins, 20% free cash flow margins, and 8% GAAP operating margin on $16.8bn Q3 revenue.
Controlling Biology - Fabian Theis, Helmholtz Munich
6 Nov · 51 min
- Fabian Theis's lab now steers experiments with Lab Compass, an inverse model for reaching desired cell states.
- Biohub announced a $1.8bn collaboration with DOE, NIH, Meta and Google for open-source biological data.
⚡ Behind Latent Space: A Rare Look Inside the Pod | AIEi Shanghai
4 Nov · 1 h 12 min
- AIE Shanghai is the first AI Engineer conference in China, brought by local partner Sana Partners.
- AIE's three-day event tops its 6,000 target with over 7,000 attendees and 1.5m online viewers.
⚡️ The 2026 Software Factory Standard — Zach Lloyd, Warp.dev
3 Nov · 35 min
- Zach Lloyd says most companies mistake heavy agent use for a software factory, which he defines as an agentic cloud system defined in code.
- Warp cuts its cost per PR from ~$80 to $15-20 and lifts throughput 40% without more reverts.
Building Legal Agents When You Can't See the Data — Jacob Lauritzen, CTO of Legora
2 Nov · 1 h 7 min
- Legora builds legal agents without seeing customer data, so a six-person eval team approximates user behavior and tracks known failure modes.
- Legora's citation accuracy jumps from 70% to 99% with a model release, with no work from the company.
Using Agents for Long-Horizon Financial Work — Aman Singh, Arca
31 Oct · 26 min
- Arca, an AI-native wealth management firm that just raised $64m, says agents replace the work around its advisers, not the advisers themselves.
- Aman Singh says Arca moved from Gemini to OpenAI to Anthropic models, with Claude as primary, after Llama proved too slow and costly despite benchmarks.
⚡ Building ChatGPT for Finance — Joseph Kim, OpenAI's Head of Financial Services
31 Oct · 28 min
- OpenAI's Joseph Kim says the new ChatGPT for Financial Services layers firm-specific templates and data sources like S&P onto consumer ChatGPT tools, with Morgan Stanley and Evercore as design partners.
- OpenAI brought financial data in house after testing big-name MCP connectors that showed error rates of 40% or more, because controlling the experience end to end was the only reliable fix.