Thariq Shihipar, a former Claude Code engineer who left Anthropic this month , posted this on X: “There’s 100% a kind of mania you can get into with coding agents, where it feels like you’re getting a lot done and working all the time but nothing is really happening.”
He added: “I think it’s often a form of avoidance of something else in life.” It is an honest line from someone who worked close to these tools. And it describes something many teams will recognise.
Questions this article answers
SAQ: the questions you should be asking (and nobody is)
Feeling productive is not being productive.
The video for this article is in production.
What he actually said
Two sentences, no study, no numbers. Shihipar did not say agents are useless. He said a certain way of using them can feel like progress without being progress . India Today reported the post on 30 October and noted he did not give a public reason for leaving Anthropic .
He is not alone in describing the pull. India Today also recalled Andrej Karpathy saying on the No Priors podcast that he was in a state of psychosis while working with agents: “I’m just like in the state of psychosis of trying to figure out what’s possible, trying to push it to the limit” .
Developer Peter Steinberger put the other side plainly in an interview quoted by The Deep View: “If you don’t ask the right questions, it’s still going to be slop.”
Activity is not output
Microsoft has been testing this inside its own company. Kathleen Hogan, its chief strategy and transformation officer, wrote that the company “learned that access and usage do not equal transformation” .
Adding agents to a broken process still leaves a broken process
Kathleen Hogan, Microsoft
She added that “speeding up one step just creates a longer queue at the next” . In other words, more agents running is a measure of activity. It says nothing yet about what got finished.
So what should you count?
Microsoft’s own answer was to start from the business outcome rather than from the tool, and to track leading indicators until results show up . For an engineer, Hogan wrote, the measure “is not how much code AI produces, but whether the team is building better products faster” .
The business lesson
Sources
Sources are listed in their original language.
- 1India Today, Armaan Agarwal, “Ex-Anthropic engineer says using AI agents may be your way of avoiding something in life,” 30 October 2026. https://www.indiatoday.in/technology/news/story/ex-anthropic-engineer-ai-agent-mania-avoidance-thariq-shihipar-3028223-2026-10-30
- 2Thariq Shihipar (@trq212), post on X, October 2026 (text as mirrored by daily.dev, accessed 30 October 2026). https://x.com/trq212/status/2115937378674372783
- 3The Deep View, Jason Hiner, “Why AI agent slop is overwhelming workers,” 2026. https://www.thedeepview.com/newsletter/why-ai-agent-slop-is-overwhelming-workers
- 4Microsoft, Kathleen Hogan, “What we’ve learned from Microsoft’s own AI transformation,” The Official Microsoft Blog, 17 September 2026. https://blogs.microsoft.com/blog/2026/09/17/what-weve-learned-from-microsofts-own-ai-transformation/