The solution might be cancelling my AI subscription

2026-06-15 · source · source ↗ · raw capture · wiki page

David Wilson's confessional-critique of AI-assisted programming from 31 May 2026, updated 3 June. Published on thoughts.hmmz.org. Links to a Cal Newport YouTube video on pseudo-productivity.

Summary

Wilson catalogues ~17 specific AI-assisted projects plus "maybe 50 others" already deleted — a Rust speech recognition system, a Jellyfin desktop clone, an Invidious clone, a regional news scraper-site, a 3D car game, an investment backtester, and many more. His assessment: "almost none of this is useful and I don't want to maintain any of it." Each session began with "write a quick script for X" and spiraled into something far larger without solving the original problem.

Key Arguments

  1. AI as attention destroyer: "A thermonuclear ADHD amplifier." Friends run three screens simultaneously on unrelated, unmaintainable projects with "so little commitment to the outcome that the time is obviously wasted."

  2. The marketing gap: Engineers build impressive things but have no plan to market them — same as before LLMs. Building for its own sake.

  3. Vendor incentives work against focus: "Almost every vendor and every tool intends to do exactly the opposite: more usage, more tokens, more output." ChatGPT is "hard-wired to include a relevant follow-up question" to drive overuse.

  4. Friction, focus, and quality are linked: His voice-to-blog experiment produced "unbridled garbage." "Because the effort was removed, so was the commitment, and with the commitment the focus, and with the focus any meaningful product at all." He argues "for as long as quality matters, handwriting can never be obsolete."

  5. Pseudo-productivity trap: Channels Cal Newport's "digital productivity paradox" — tools make individual tasks faster but leave workers busier, more distracted, and less productive overall. AI users spend more time in shallow tasks and less in focused work.

Conclusion

Wilson concludes a tool delivering "a cheap reward with minimal input and no friction can only be a liability." He has no solution beyond curtailing use, and notes that realizing this "is probably the only real contribution of AI to date."

Themes

Attention economics, pseudo-productivity, friction and meaning, the dark side of lowered creation barriers, engineering culture, sustainability of output.