Greg Kamradt and Daniel Mac debate AGI: 'general' means able to learn anything, not be trained
GregKamradt · x · 2026-09-09
- Daniel Mac proposed inverting the AGI definition: "general means a machine mind that can be trained to do anything," since humans — the existence proof of general intelligence — can't do many things without training.
- Greg Kamradt (ARC Prize) agreed learning is the key but suggested "a machine mind that can learn to do anything," dropping "trained" because it implies an external trainer rather than an inherent capability.
- Consensus: generality hinges on learning itself, not on untrained performance.
Related event: ARC Prize Founder and Community Debates What "General" in AGI Means(5 posts)→
More from AGI Musings
- Eric Topol challenges frontier-model study: where's proof of sustained upskilling? — EricTopol · 2026-10-27
- Deep dive: where do p(doom) estimates of AI apocalypse actually come from? — anderssandberg · 2026-10-27
- Pedro Domingos: as intelligence gets cheap, demand will skyrocket — pmddomingos · 2026-10-27
- Apollo chief economist: AI has delivered zero financial payoff outside Big Tech — rseroter · 2026-10-27
- OpenAI acquires Orchestra Bio to bring clinical-trial monitoring agents into drug development — dl_weekly · 2026-10-27
- Do AI agents work better when they pace themselves and take breaks? — EpistemicHope · 2026-10-27