Re:The "confident idiot" problem: Why AI needs hard rules, n...
by LogicFailsMe on Monday December 08, 2025 @01:04PM
(#46195554)
I am reminded of AI summaries and Microsoft Copilot. All push low value. But I separate that from the underlying potential of the technology. And I wish we heard more from deep domain experts like Karpathy and less from influencer dilettantes like Dylan Patel about where this is going.
Re:The "confident idiot" problem: Why AI needs hard rules, n...
by thfuran on Monday December 08, 2025 @02:28PM
(#46196523)
I want to query a bayesian ontology, not a Markov chain with delusions of grandeur.
Re:The "confident idiot" problem: Why AI needs hard rules, n...
by TeMPOraL on Monday December 08, 2025 @04:18PM
(#46197771)
Alas, computation costs energy, so you get what you can afford.
Also one thing I thought LLMs did already is kill the misguided idea of applying prescriptive, formal categorization to the real world.
And once this garbage is in your context, it's polluting everything that comes after. If they don't know, I need them to shut up. But they don't know when they don't know. They don't know shit.