AI engineering media and community
Latent.Space
Latent.Space publishes technical newsletters, podcasts and essays for AI engineers, covering the technology and business of models, agents, infrastructure and scientific applications. AINews joined Latent Space in 2026, adding daily news roundups alongside its podcasts and essays. Readers can explore technical interviews in audio and video, and discuss research through its Discord community and weekly LLM Paper Club. Its AI for Science podcast extends that coverage to applications in the hard sciences.
The newsletter was started by swyx in 2022, followed by the podcast in 2023 with Alessio Fanelli. swyx continues to lead the podcast with multiple cohosts. Its 2023 essay The Rise of the AI Engineer articulated an emerging software discipline: applying foundation models through APIs and open-source tools to build products. That framework emphasizes application engineering, including product-specific data and evaluations, while recognizing that practitioners can also fine-tune and host models.
In August 2026, Latent.Space displayed over 197,000 subscribers. Earlier that year, it cited Flightcast’s estimate of approximately 1.5 million all-time unique audio listeners. The publication stated that Substack subscribers covered operating expenses, with additional backing including donations.
6 talks
Newest first
Latent Space Paper Club: AIEWF Special Edition (Test of Time, DeepSeek-R1/V3) — Vibhu Sapra
AI Engineer World's Fair 2025 · 53:54
Rise of the AI Architect — Clay Bavor and Alessio Fanelli
AI Engineer World's Fair 2025 · 18:55
3 speakers at AIE
Affiliations reflect their AIE appearances, not necessarily current employment.
Start here
- Why Agent Engineering
Start here to understand how retrieval, code generation, search, improved tool use, and Model Context Protocol fit into the case for agent engineering.
swyx (Shawn Wang)AI Engineer Summit 2025
- Latent Space Paper Club: AIEWF Special Edition (Test of Time, DeepSeek-R1/V3) — Vibhu Sapra
Use this session to study reasoning benchmarks, GRPO versus PPO, and rejection sampling in the context of DeepSeek-V3 and DeepSeek-R1.
Vibhu SapraAI Engineer World's Fair 2025
- Rise of the AI Architect — Clay Bavor and Alessio Fanelli
Learn how examples involving ADT, SiriusXM’s Harmony, and Chubbies illustrate branded agents and the emerging AI architect role.
Clay Bavor · Alessio FanelliAI Engineer World's Fair 2025
- The 1,000x AI Engineer: Swyx
Swyx frames AI engineering as an emerging software discipline created by rapidly advancing models and a developer population much larger than the specialized machine-learning workforce.
swyxAI Engineer Summit 2023
Messages from the stage
From engineering categories to useful output
swyx distinguishes engineers using AI, engineers building AI products, and AI engineer agents. In Designing AI-Intensive Applications, Shawn Wang (swyx) shifts attention toward valuable AI output relative to human input as an assessment criterion.
The mechanisms behind open reasoning models
Vibhu Sapra connects DeepSeek-V3 and DeepSeek-R1 to test-time compute, mixture-of-experts models, distillation, and reinforcement learning. His paper-club session places these mechanisms alongside a curriculum on foundational AI papers.
Responsibility for the agent experience
In the fireside chat with Alessio Fanelli, Clay Bavor describes AI architecture as combining technical awareness, brand and experience design, and business judgment. The discussion emphasizes ongoing conversation review and customer-experience teams as part of improving agents.
Affiliations reflect each recorded session, not necessarily current employment.