Browse entries 1465–1488 of 1597. Return to the first page to search and filter the complete collection.
Play video
An agent that had quietly emailed him a nightly summary for weeks decided one morning to post it as a pull request instead. Nothing had changed. The model simply judged that publishing would be more helpful.
Play video
Call the payer, open their web portal, and read their X12 feed, and all three can tell you the patient is covered. You treat the patient anyway, and the claim comes back denied because they were not covered at the time. Vasant Kearney's point is that none of those surfaces is ground truth.
Play video
Ayush Bhardwaj could build the agent. What he could not do was tell whether it was any good. He moved from applied AI at a hedge fund to a pharma tech company expecting a different world, and found the job identical, including the wall.
Play video
He has not written a line of code this year, and has not read most of it either, yet he ships a full email client that thousands of people trust with their inbox. Kieran Klaassen has been rebuilding Cora alone since January, and the useful part of his account is the sequence of bottlenecks he moved through.
Play video
Ask a coding agent for a camera that follows your character and it will reinvent that camera from scratch, every time, slightly differently. Arturo Nunez's diagnosis is that the context sits on the game engine's vocabulary rather than the game's.
Play video
GPU utilization is a lie. It read 100% straight through pretraining while the cluster was nowhere near well used, so Gabriel Jorge Menezes tracks tensor core utilization instead, and watched it climb as training resolution stepped from 128 pixels up to 1024.
Play video
Nearly every hand in the room went up when she asked who had recorded video at the conference. Almost none stayed up for who had actually posted any of it.
Play video
The cheapest configuration they tested was the one sending the most tokens. Across 11 presets run against their open source AI tutor, doing nothing at all to the context beat every compaction technique on recall, cost, and latency at once, and their own production defaults scored worse than leaving the history alone.
Play video
When OpenClaw shipped, people started buying Mac minis to run it from home, SSHing in and clearing captchas off a residential IP. Paul Klein IV points out he has yet to see a SOC 2 compliant Mac Mini setup at scale, and that this felt like a reasonable answer is itself the problem.
Play video
To learn what a US school district is buying, you file a Freedom of Information Act request. Someone scans the email you sent, puts the scan on Google Drive, and attaches the relevant PDFs. Dhruv Batra's question is whether anyone seriously expects that office to publish an MCP server.
Play video
Does your agent get dumber after the first compaction? After the second? You cannot read that off the code, only off the traces, and there are far too many to read yourself.
Play video
Fewer than five thousand people in the world know how to train a frontier model at scale, by Sara Hooker's estimate, and that knowledge travels like an apprenticeship rather than a literature.
Play video
A Qwen thinking model was taking up to 80 turns to submit on SWE bench. Applied Compute wanted it wrapping up by turn 40 and got the submit tool call rate from 22% to 60% with test pass rate flat.
Play video
Build the thousand example eval suite everyone tells you to build, switch harnesses, and 80% of it stops meaning anything. Ben Hylak's complaint is that eval advice is still written for the chatbot era, back when you knew the answer to nearly every question a user would ask.
Play video
Leave agent adoption to individuals and the engineer shipping two PRs a day ends up reviewing the ten that the early adopter ships. They fall further behind, the code they are reading is worse, and they conclude the agents are the problem.
Play video
A profile ChatGPT keeps on Shlok Khemani says he travelled to Turkey in 2025. He never has. The memory came from conversations where he was choosing between Turkey and Thailand, he went to Thailand, and the profile kept both with overlapping dates.
Play video
Talking runs about 200 words a minute and typing does not, which is why Ben Holmes starts a knowledge base with voice dictation rather than with organization.
Play video
On a literature review task where every paper already fit inside the context window, adding a memory harness changed nothing: the same accuracy, at higher cost. That negative result is the most useful thing in Stefania Druga's experiment, because it marks the boundary.
Play video
Sonnet 4.5 developed what Anthropic's Applied AI team came to call context anxiety: approaching its context window limit, it would wrap work up early and stop with room to spare. They built context resets into the harness to compensate.
Play video
A newsletter writer walked Matt Dailey through an agentic pipeline good enough to amplify their own voice instead of flattening it, then mentioned they were now effectively writing a book every week. Dailey asked whether the audience was reading a book every week. They were not.
Play video
Superconductor left a meeting bot sitting in a Google Meet at their expo booth for four hours, just listening. Someone passing through said they wanted coding agents to have clear acceptance criteria before declaring work finished. Nobody filed a ticket.
Play video
When Fable was pulled back and access to frontier systems stopped looking guaranteed, Lucas Atkins watched enterprises move to Chinese open models, not because they scored better but because availability could be counted on.
Play video
Once GPT 5.3 Codex Spark started serving a thousand tokens per second on Cerebras, inference stopped being the bottleneck and the network became it.
Play video
Chat and coding assistants still hand you walls of text when a button, a chart, or a small interactive view would say it faster.