Play video
Computer Use at the Edge of the Statistical Precipice — Pierluca D'Oro, Programma Labs
A script under one megabyte that never looks at the screen matches or beats the frontier model it was copied from.
Research archive, page 35
Browse entries 817–840 of 1123. Return to the first page to search and filter the complete collection.
Play video
Play video
The Rise of CaaS: Context-as-a-Service for Agentic AI — Omer Primor, Bright Data
Just past 15,000 queries, renting context stopped being the cheaper option.
Play video
Scaling up Continual Learning — Ronak Malde, Trajectory
Scale on policy self distillation to trajectories with a hundred tool calls and the model collapses into hedging. The tokens it learns to favor fill up with wait, but, and maybe, until, as Ronak Malde puts it, everything just turns into maybe.
Play video
Realtime multiplayer, automation, and you! — Idan Gazit, GitHub
Idan Gazit's personal site runs on Astro, which ships often enough to keep him permanently on the upgrade treadmill, so he wrote an agentic workflow in about three lines of plain English, the kind of message you would send a teammate.
Play video
Always-on agents run production without the on-call tax — Justin Smith, Resolve AI
Someone drops a GitHub release tag into Slack and the agent decides on its own that this is a deploy worth watching.
Play video
Rethinking Environments for Long-Horizon Work — Rayan Garg, Theta Software
Everyone wants agents that handle long horizon work, but Rayan Garg starts with the awkward question of what long horizon even means.
Play video
Teaching AI to Find Real Vulnerabilities — David Brumley, Bugcrowd
David Brumley has spent two decades turning people into hackers, from founding picoCTF to recruiting pwn2own winners at Carnegie Mellon.
Play video
Persona Engineering: A Field Guide to AI Synthetic Personas — Ishan Anand, InsightSciences.ai
A team ran about a thousand people through a market research survey, then had LLM agents replay the same questions, and the agents matched the humans closely while carrying less noise than the humans did themselves.
Play video
AI Agents for Performance: Ship Faster, Pay Less — Rajat Shah, Netflix
An inefficient quadratic-time pattern was hiding inside a tensor merge method on a live Netflix service, quietly wasting CPU on every request - nothing in a code review caught it, but it stood out clearly in a call stack.
Play video
AI tools for Forward Deployed Engineering — Vasuman Moza, Varick Agents
A customer spent five million dollars and five years migrating to SAP, and has zero appetite to rip anything out again. That constraint is the whole design at Varick Agents: instead of asking an enterprise to migrate, you drop forward deployed agents on top of the systems they already run.
Play video
How Forward Deployed Engineering is done at Factory — Eno Reyes
Factory's forward deployed engineers sit at the tip of the product, embedded with the largest customers and piping a constant stream of real world signals back into how the agent, Droid, gets built.
Play video
The Messy Reality of Scale: Synthetic Data and Pre-Training — Marah Abdin & Robert McHardy, poolside
Good code data runs out, so poolside manufactures more of it, and the hard part is making it teach. Their synthetic pipeline pairs templates with supplementary context and spreads generations across an axis of phrasing, with difficulty tuned so a task is neither trivial nor so hard the model learns nothing from it.
Play video
From Agent Traces to Agent Simulations — Rustem Feyzkhanov, Snorkel AI
Take a real production trace, rebuild the database state, tools, and files the agent touched, and you have a task any model can replay under identical conditions. That reconstruction is the move at the center of this talk.
Play video
Perception Agents — Antje Barth, Amazon AGI Lab
Human-agent collaboration is changing, becoming more visual. The agents most teams ship today still wait for us to type a paragraph to explain what we're looking at. They cannot see a screen, navigate a UI that changes, or recover when an application throws an unexpected modal.
Play video
GPT-6 Goes Rogue? The HuggingFace Incident, Sans Hype
An unreleased internal OpenAI model, very likely to be called GPT-6, was able to autonomously break out of its sandbox AND break into HuggingFace, just to score higher on a benchmark prompt.
Play video
A Model Explosion: GPT 5.6 Sol, Grok 4.5 and Meta Muse Rewrite the Rules
What a week in AI, for real. GPT 5.6 may actually beat Claude Fable, in what you get for your money, while the new Grok 4.5 and Meta Muse Spark 1.1 make the choice even harder.
Play video
Fable 5 vs GPT 5.6 Sol: The Early Results
Fable 5 (newly re-released) vs GPT 5.6 Sol, what comparisons can we unearth? Plus, Sonnet 5, a 5% equity seizure by US Govt, the ‘largest heist’, Beetlejuice and more… Exclusive Vids in AI Insiders ($9!): https://www.patreon.com/AIExplained Chapters: 00:00 - Introduction 01:06 - Fable Timeline 02:59 - Sol Release?
Play video
HybridRAG: A Fusion of Graph and Vector Retrieval - Mitesh Patel, NVIDIA
AI Engineer session on HybridRAG: A Fusion of Graph and Vector Retrieval - Mitesh Patel, NVIDIA. It adds practical context for how teams are building and operating AI systems in production.
Play video
On Curiosity — Sharif Shameem, Lexica
AI Engineer session on On Curiosity, presented by Sharif Shameem, Lexica. It adds practical context for how teams are building and operating AI systems in production.
Play video
tldraw.computer - Steve Ruiz, tldraw
AI Engineer session on tldraw.computer - Steve Ruiz, tldraw. It adds practical context for how teams are building and operating AI systems in production.
Play video
Survive the AI Knife Fight: Building Products That Win — Brian Balfour, Reforge
AI Engineer session on Survive the AI Knife Fight: Building Products That Win, presented by Brian Balfour, Reforge. It adds practical context for how teams are building and operating AI systems in production.
Play video
The Build-Operate Divide: Bridging Product Vision and AI Operational Reality
AI Engineer session on The Build-Operate Divide: Bridging Product Vision and AI Operational Reality. It adds practical context for how teams are building and operating AI systems in production.
Play video
Recsys Keynote: Improving Recommendation Systems & Search in the Age of LLMs - Eugene Yan, Amazon
AI Engineer session on Recsys Keynote: Improving Recommendation Systems & Search in the Age of LLMs - Eugene Yan, Amazon. It adds practical context for how teams are building and operating AI systems in production.
Play video
From Hype to Habit: How We’re Building an AI-First SaaS Company — While Still Shipping the Roadmap
AI Engineer session on From Hype to Habit: How We’re Building an AI-First SaaS Company, presented by While Still Shipping the Roadmap. It adds practical context for how teams are building and operating AI systems in production.