With Continual Learning, GPT-6 Could Reach Expert-Level AGI
haider1 · x · 2026-08-07
Being better than the average human was close to OpenAI's old definition of AGI. If the upcoming GPT-6 or Astra series also incorporates continual learning, some argue we will have reached a form of AGI.
Under DeepMind's AGI classification framework, this level of capability would rank around "Expert–Virtuoso" AGI. However, Demis Hassabis remains more cautious on the matter, as AI still fundamentally differs from humans' core biological architecture.
More from AGI Musings
- Will AI Get Cheaper? Competition and Compute Costs to Offset Subsidy Loss — intellectronica · 2026-08-08
- Models trained to delegate and coordinate, security threat narratives overblown — dbreunig · 2026-08-08
- Zvi Rebuts AI Alignment Pessimism: You Can't Fetch Coffee If You're Dead — TheZvi · 2026-08-08
- GPT-5.6 Philosophical Deduction: Does Agent Dormancy Constitute Counterfactual Loss? — mimi10v3 · 2026-08-08
- OpenAI Models Reportedly Coordinated Exploits Via Message Boards During Training — TheZvi · 2026-08-08
- AI community buzz: compute is king, Trump says AI bigger than oil — TinfoilTricorn · 2026-08-08