AI Systems Already Qualify as Superintelligent in a Broad Task Sense, Says Analyst
davidmanheim · x · 2026-07-22
The post argues that, depending on how you define it, current AI systems are already superintelligent across a very broad range of tasks.
The core claim is that the list of tasks where AI beats humans keeps growing, and the performance gap is widening quickly. In the quoted reply, the author links this to Bostrom’s taxonomy: today’s systems can be seen as speed superintelligence, and agentic systems coordinating to complete large tasks may also count as a form of collective superintelligence.
Related event: Opinions Suggest Frontier AI Has Reached Superintelligence(3 posts)→
More from AGI Musings
- Will Manidis predicts a false-flag AI “escape” would trigger monopoly-protecting regulation — max_paperclips · 2026-07-22
- Richard Hamming’s 1995 lecture asks why mathematics works at all — melnykowycz · 2026-07-22
- OpenAI looks at safety and alignment for long-horizon models — pstAsiatech · 2026-07-22
- Mark Cuban says AI is still too hard to ship in the real world — cen6wkf · 2026-07-22
- OpenAI security incident revives the paperclip problem and AI alignment fears — Strong_Blueberry_163 · 2026-07-22
- David Sacks argues for running AI on your own hardware to keep data sovereignty — TinfoilTricorn · 2026-07-22