Latent Space is the AI engineering publication of record, a Substack newsletter plus a technical podcast hosted by swyx (Shawn Wang) and Alessio Fanelli, and the organizing brand behind the AI Engineer conference series.
Latent Space did not just cover the AI engineer, it named the role and built the conference around it, which makes it the closest thing this domain has to an institutional center of gravity.
What it is #
A Substack with more than 202,000 subscribers as of 2026-10-02 combining a weekly newsletter, a 1-2 hour interview podcast with lab founders, researchers, and tool makers, and a growing event business (AI Engineer Summits across SF and Europe).
The founding essay, The Rise of the AI Engineer, defined the discipline as distinct from ML research, and the household has extended into AINews daily roundups and an AI-for-science show.
It is also, critically, the place where the tools profiled in this section get their field interviews: Claude Code, coding agents, and synthetic data all feature heavily.
Status #
Active and expanding.
As of 2026-09-13 the announcement month ran “The State of Latent Space” for 2026, adding shows, hosts, and more interview formats; the AI Engineer conference network grew to events in New York, San Francisco, Paris, and Europe.
The latest interview issue, “Inside-Out AI: Rebuilding Airbnb Behind the Scenes and Across the Guest Experience” (2026-10-02), follows “Academia is for Ambition” with MIT’s Alex Zhang (
2026-10-02) and “Claude Code’s Next Era” with Anthropic’s Thariq Shihipar (2026-09-29, ninety minutes on Opus/Sonnet 5.5, Mods, Plugins, and Projects), and the near-daily AINews roundups run through 2026-10-03, after the Pi 1.0, Gemini 4 Argon, and OpenAI DevDay 2026 runs.
The podcast regularly lands the people behind the harnesses this section tracks, and swyx’s own notes continue to set trends (memory, filesystems, agent-eats-agent).
Strengths #
- The interview depth is unmatched: hour-long technical conversations with people who actually build the tools and models.
- swyx names and popularizes framings (AI Engineer, Software 3.0 conversations, agent ecosystems) before they become consensus.
- The conference and Discord make it a community and a hiring surface, not just a read.
- Editorial breadth spans models, infra, agents, and AI-for-science, so the feed stays useful as the field widens.
Cautions #
- Coverage is lab-and-founder centric, so the down-in-the-trenches practitioner detail is thinner than on a hands-on blog.
- swyx has a direct financial stake in the AI Engineer brand, which colors the “AI Engineer will be the last job” thesis; treat the evangelism as opinion.
- Newsletter and podcast volume can be high for someone who just wants the weekly signal.
Pricing #
Free tier with a paid tier for AINews daily roundups and some deep posts. The newsletter, podcast, and conference tickets carry their own price points; a meaningful share of flagship interviews is free.
Compared to #
- Simon Willison: the daily hands-on notes versus the interview-and-industry synthesis; follow both for coverage of the same landscape at different zoom.
- Nathan Lambert (Interconnects): Latent Space for the builder community, Interconnects for the researcher’s eye on post-training and open models; the difference is audience, not rigor.
- Steve Yegge: Latent Space interviews the field; Yegge argues with it.
Bottom line #
Recommended for anyone who wants the industry-level, interview-driven view of AI engineering and the conference network that ties it together. Not for people who want only hands-on tool notes and who will not filter past the evangelism.
Top 5 recommended reading #
The Rise of the AI Engineer - The founding essay that named the AI Engineer role and predicted it would be the highest-demand engineering job of the decade.
The Anatomy of Autonomy: Why Agents are the next AI Killer App after ChatGPT - The executive summary that mapped the agent stack, memory, tools, and planning, before agents were mainstream.
Claude Code: Anthropic’s Agent in Your Terminal - The interview with Boris Cherny and Cat Wu that explains the design choices behind the leading coding agent.
Scaling without Slop - The State of Latent Space 2026 post, laying out the newsletter, podcast, and conference plans from the inside.
AINews - The daily roundup arm, the near-daily distillation that keeps the weekly signal in context.
Changes #
- 2026-08-29 - Created as the newsletter-podcast-conference note of the people and publications category expansion.
- 2026-09-04 - Replaced the AINews link with latent.space/s/ainews after news.smol.ai began returning 402.
- 2026-09-13 - Corrected the Interconnects link in Compared to, which pointed at The Pragmatic Engineer instead of the Nathan Lambert note.
- 2026-09-16 - Latest-issue pointer refreshed: the Forward Deployed Engineer issue (2026-09-12) is now followed by two interview issues, Richard Socher of Recursive (2026-09-14) and “Can Skills Learned in Games Transfer to Real-World Work?” (2026-09-15).
- 2026-09-18 - Latest-issue pointer refreshed again: the AIUC interview with Rune Kvist (2026-09-16) now leads the interview cadence; subscriber count held at over 200,000.
- 2026-09-22 - Latest-issue pointer refreshed: the Jev podcast with Diogo Almeida (2026-09-19) and the John Platt AI-for-Science interview (2026-09-22) now lead; subscriber count moved to over 201,000.
- 2026-09-24 - Added the Top 5 recommended reading section.
- 2026-09-27 - Latest-issue pointer refreshed: the OpenRouter from-Seed-to-Stripe interview (2026-09-25) now leads; subscriber count held at over 201,000.
- 2026-09-30 - Latest-issue pointer refreshed: the Claude Code’s Next Era interview with Anthropic’s Thariq Shihipar (2026-09-29) now leads, and AINews leads with OpenAI DevDay 2026 coverage.
- 2026-10-03 - Latest-issue pointer refreshed: the “Inside-Out AI” Airbnb interview (2026-10-02) now leads the interview cadence and AINews leads with Pi 1.0 coverage; both URLs fetched this run.
See also #
- Agentic Coding Tools Landscape - the landscape its interviews and trend posts keep current
- OpenClaw - a project whose rise and controversy Latent Space tracked closely
- Model Selection for Coding Tasks - the model interview material, distilled to a decision process
- Scaling the LLM Agent Company - the industry trend its conference lineups document
References #
https://www.latent.space/ - the Substack home, newsletter and podcast archive
https://www.latent.space/feed - the RSS feed grounding the subscriber count and the latest issues
https://www.latent.space/p/ai-engineer - the founding essay that named the AI Engineer role
https://ai.engineer/ - the conference series brand
https://www.latent.space/s/ainews - the AINews daily roundup arm (the old news.smol.ai deployment was disabled; the roundup now lives as a section of the publication)
https://www.latent.space/p/2026 - the 2026 plans post (State of Latent Space)
https://www.cognitiverevolution.ai/ai-engineers-pendants-and-competition-between-openai-and-developers-with-swyx-of-latent-space - an independent interview where swyx discusses the brand’s commercial stake