David Pierce, editor-at-large at The Verge and co-host of The Vergecast, estimates he has installed, used, and abandoned roughly 200 to-do apps over 20 years of testing productivity software. In a new interview with Casey Newton on Platformer, published July 14, 2026, Pierce delivers a blunt verdict on the current AI moment: the tools are useful for busywork, unreliable for judgment, and nowhere close to the revolution being sold.
Pierce is arguably the tech press's most-cited productivity correspondent, with prior stints at the New York Times alongside David Pogue, Wired, the Wall Street Journal, and Protocol. His weekly newsletter Installer and his show Version History cover what people should actually download and use. That makes his skepticism notable — this is the person paid to try everything.
His frame for the last 18 months is a 'chaos phase.' Everyone has decided everything is about to change, he told Newton, but none of it actually has yet. The pitch that workers must adopt AI or fall behind, in his read, is a marketing frame that has outrun the product.
“We're in a moment where everyone has decided that everything about how we do everything is about to change. But actually, none of it has changed yet.”— David Pierce, Editor-at-large, The Verge
Key facts
- 01David Pierce estimates he has installed, used, and abandoned roughly 200 to-do apps, limited only by the number that exist.
- 02Pierce clocks AI clearing a thousand marketing emails from his inbox in 45 seconds — but says it got 'almost everything wrong' on his primary inbox.
- 03He cycles between five task apps, including Todoist and Apple Reminders, and four or five notes apps depending on the pain point of the day.
- 04Pierce frames the last 18 months as a 'chaos phase' where AI hype has outpaced any real change in how work gets done.
- 05His advice: find-and-replace the word 'AI' with 'software' whenever the hype feels overwhelming.
Where AI does deliver, per Pierce, is the low-judgment busywork of daily life. He points to converting a folder of files, finding a plumber, or clearing a thousand marketing emails from an inbox in 45 seconds as tasks the current generation of tools handles well. These are real productivity wins, and he doesn't dismiss them.
The failure mode shows up when the work requires taste. When Pierce pointed AI at his primary inbox, it started getting almost everything wrong — fooled, he suspects, by senders who write every message as a five-alarm fire. He reads this as the leading edge of a new attention-hacking pattern that AI triage tools will need to solve before they can be trusted with prioritization.
The advice Pierce offers people who feel behind is unusually direct for someone whose job is covering new products. Offload the dumb tasks, keep doing the work that builds your own understanding, and when the hype feels heavy, mentally swap 'AI' for 'software.'
On his own setup, Pierce is candid that nothing has stuck and the arrangement is incredibly fragile. He cycles between roughly five task apps, switching between Todoist and Apple Reminders depending on the day, and bounces between four or five notes apps. He switched notes apps the morning of the recording — leaving NotePlan for Craft after a Craft update improved task integration and added Apple Reminders sync.
His operating theory is that capture is the most important thing by a mile, and none of the rest of it matters. That's the argument that most cleanly connects his productivity view to the AI moment: if models can organize and retrieve unstructured input after the fact, elaborate tagging systems become dead weight. Apple Reminders wins his loyalty because Siri integration makes capture nearly friction-free, even though Todoist is a better project manager.
The counterweight to Pierce's view is that the frontier is moving. Agentic systems from Anthropic — where, Newton discloses, his fiancé works — and other labs are being sold specifically on the promise of exercising judgment across long tasks, not just executing busywork. If those agents mature, the gap Pierce identifies between 'good at bulk' and 'bad at priority' narrows, and the find-and-replace 'AI is just software' framing gets harder to sustain.
For AI Chat Daily readers, Pierce's frame is a useful corrective to the productivity-tool marketing cycle. The AI productivity market is currently valued on the premise that everyone must adopt or lose ground, which sustains valuations at labs and app-layer startups alike. A working reporter who tests everything saying the returns are modest, the judgment layer is broken, and staying ahead is overrated is not a bearish call on AI — it's a bearish call on the pricing power of AI-branded productivity software specifically. The apps that will survive this cycle are the ones that make capture faster, not the ones that promise to replace the human doing the prioritizing.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
