This is a problem because in the high-stakes applications such as medicine, engineering, and scientific research, it matters not only what a system concludes but also how it arrives at its conclusion. When mistakes happen — for example, in medical diagnosis and treatment — we need to be able to pinpoint what went wrong: Was the system’s reasoning at fault, did it draw on invalid evidence, or did it make incorrect assumptions?
Search
5 results for “timokissel”
-
And "chain of thought" isn't reasoning. It is still produced by the same next-token prediction process, just iterated for longer.
-
Today’s #AI lacks genuine reasoning. If we want future AI systems to produce trustworthy results and really novel insights in fields like science and medicine, we need to equip them with genuine reasoning capabilities, and old-school stuff like tree search, expert systems, logic solvers, etc. will come in handy.
https://www.technologyreview.com/2026/10/02/1145639/dont-be-fooled-llms-dont-reason/
-
"Safety only makes things slower when it’s tacked on, outside of the core technology"
Sadly, even though we know quite a bit of how to solve such issues, we aren't moving #technology forward at pace if we don't approach this from an integrated perspective early.And, as I keep reminding people, the #friction is the point. It's a feature, not a bug.
-
Two key quotes here:
"The human just becomes this meat tool to give permissions without the cognitive capability to engage"
This is the trap if your "value-add" consists of solely typing questions into a #chatbot. Perhaps ironically, this will make people who lack that #AI "skill" but can solve things on their own even more valuable.#productivity
https://spectrum.ieee.org/agentic-ai-humans-in-loop