In 30 seconds
Theo introduces Andrej Karpathy's discussion of the Claude tagging feature in Slack. Although he finds the post awkward and jokes about having to defend Anthropic and discuss Slack, he thinks the underlying idea is interesting and has lessons for the industry.
- 1An unexpected subjectTheo says it may be surprising to see a prominent AI researcher discussing a Slack bot.
- 2A reluctant defenseHe jokes that he must defend Anthropic, make a video and talk about Slack.
- 3An idea worth examiningDespite his reaction to the post, Theo thinks the Claude tagging feature has lessons for the industry.
Takeaways
- Theo sees value in examining Claude's tagging feature in Slack despite skepticism about the post.
- He believes the feature raises ideas the AI industry can learn from.
- The short video introduces his reaction without explaining the feature's mechanics in detail.
Read the edited text by chapter
A surprising post about Claude in Slack
Theo introduces Karpathy's discussion of a Claude tagging feature. He finds the subject surprising but thinks there is something valuable behind it.
If I told you that Andrej Karpathy, one of the prominent researchers who helped shape today's language models, was talking about a Slack bot, you might think either he or I had lost the plot. I understand the skepticism.
That leaves me doing three things I don't enjoy: defending Anthropic, making a video about it and, worst of all, talking about Slack. But however awkward I find the post, I think what Karpathy is discussing about the Claude tagging feature contains important lessons for the industry. I actually think it's pretty cool.
More from this channel
38:47EN
Theo - t3․gg· Oct 28, 2026Theo’s Long-Running Agent Experiments: Verification, Compute and CostTheo Browne · Ping LabsA skateboard game and two compiler projects show how repeated checks and flexible compute can sustain agent work—and how expensive those loops can become.
36:07ENTheo - t3․gg· Oct 27, 2026Theo on Token Costs, Coding Agents, and the Price of AI CapabilityTheo Browne · Ping LabsA viral estimate of Lauren's AI spending prompts a closer look at cache pricing, benchmarks, and the changing work of coding agents.
30:37ENTheo - t3․gg· Oct 26, 2026Theo’s Three Skills for Reviewing and Shipping with Coding AgentsTheo Browne · Ping LabsHow sol-loop, full-send and full-stop encode review, verification and different levels of merge authority.
0:43ENTheo - t3․gg· Oct 23, 2026Theo Distinguishes Security Vulnerabilities from Active AttacksTheo Browne · Ping LabsA short explanation of why disclosed software flaws and reported attacks are different kinds of security news.
29:22ENTheo - t3․gg· Oct 22, 2026Theo on Security Vulnerabilities, Stolen API Keys and AITheo Browne · Ping LabsA look at the gap between vulnerability disclosure and patching, and a speculative account of why stolen inference keys may be valuable.
31:13ENTheo - t3․gg· Oct 20, 2026Theo on Why Smarter Models Help More Than DevelopersTheo Browne · Ping LabsFrom game modifications to compiler experiments, Theo revisits his assumptions about who benefits from more capable AI.