Skip to the good part of every AI Engineer talk.
Find the ten minutes you need in a forty-minute talk. 1138 talks and counting.
Search talks and packs by topic, or describe the problem you're working on.
Why AI Agents Need Million-Token ContextOlive Song presents MiniMax M3 as a model built for agents that need to retain long conversations, tool responses and information from complex environments. M3 has around 400 billion total parameters, 20 billion activated parameters, vision and video understanding, and...
Agents' Next Frontier: Agent-to-Agent and Network EffectsJean-Denis Greze argues that agent-to-agent is a misleading way to frame the problem. The real task is search: before an agent responds or calls a tool, it needs the right information in its context window. The ideal system would give...
Everyone Gets A Software CompanyBenjamin Guo presents Zo Computer as a response to software subscriptions, fragmented services, and cloud systems that keep users from owning their data and tools. He calls this arrangement technofeudalism: people pay SaaS providers, those providers pay cloud companies, and...
Coding agents look capable in a fresh repository and come unstuck in a ten-year-old one.
A production agent can give the right answer and still refund the same customer twice, act on a stale record or quietly finish half a job.
The chat box makes users package their work into prompts, wait, inspect the answer and try again.
More context can make an agent worse.
Every talk and pack page also exists as plain Markdown: add .md to any talk or pack URL.
