Every "impressive" AI use case fails one of two tests, argues Redditor
gaykenyan_lion · reddit · 2026-09-23
The author offers a two-pronged critique of "impressive" AI use cases: either the task wasn't worth doing at all (endless email summaries, automating admin that just creates overhead, slop), or doing it the hard way would have delivered a massive cognitive benefit that outsourcing to AI forfeits. A short, deliberately open-ended provocation.
More from AGI Musings
- Jagged frontier persists: models still struggle to abstract novel concepts into working patterns — Dan_Jeffries1 · 2026-09-23
- Pacing the Frontier: A Research Agenda to Curb AI Race Incentives Beyond Blunt Slowdowns — gleech · 2026-09-23
- New essay argues 'pacing the frontier' is a privilege AI can't afford as alignment lags — AccBalanced · 2026-09-23
- Toby Walsh calls out Hugging Face ethics lead over Books3 distribution and disinformation campaign — TobyWalsh · 2026-09-23
- Wired Survey: 59% of 634 Women in Tech Saw Gender Bias, Slam AI Double Standards — AnnaCiaunica · 2026-09-23
- As AI job losses loom, advice is to plan a career change proactively — shashib · 2026-09-23