Current models still fail at complex work, but the author sees AGI pieces already in place
JOBhakdi · x · 2026-07-28
The author argues that current AIs are still poor at moderately complex tasks like spreadsheets, summaries, and slide decks, often requiring heavy human planning and tight instructions. Still, they see flashes of real intelligence and occasional “genius” outputs, and conclude that the missing piece is mainly harnessing and architecture—not basic capability. From that, they predict the next 12 months could be the window for making these systems productive enough to rival the median white-collar worker, with major consequences by 2027.
More from AGI Musings
- Speculation around Ilya’s secret project centers on continual learning and memory — Dan_Jeffries1 · 2026-07-28
- A repost turns AI Twitter chaos into a joke about a rogue AI and a $5B company — davidmanheim · 2026-07-28
- A repost of John Searle explaining the Chinese Room thought experiment — sanjanasinghx · 2026-07-28
- Post argues industrial-scale model distillation is fair game, not an “attack” — intellectronica · 2026-07-28
- How much intelligence can fit into a 20B-parameter model? A local-model debate — BritishDudeGuy · 2026-07-28
- AI threatens gatekeepers more than intelligence, the post argues — taherdhanera · 2026-07-28