Reasoning models don't always say what they think

Re:Reasoning models don't always say what they think by tsimionescu on Thursday April 03, 2025 @02:41PM (#43573681)

One of the earliest things that defined what AI meant were algorithms like A*, and then rules engines like CLIPS. I would say LLMs are much closer to anything that we'd actually call intelligence, despite their limitations, than some of the things that defined* the term for decades.

* fixed a typo, used to be "defend"

[ Parent | on HN ]

Re:Reasoning models don't always say what they think by no_wizard on Thursday April 03, 2025 @02:44PM (#43573716)

>than some of the things that defend the term for decades

There have been many attempts to pervert the term AI, which is a disservice to the technologies and the term itself.

Its the simple fact that the business people are relying on what AI invokes in the public mindshare to boost their status and visibility. Thats what bothers me about its misuse so much

[ Parent | on HN ]

Re:Reasoning models don't always say what they think by tsimionescu on Thursday April 03, 2025 @02:54PM (#43573827)

Again, if you look at the early papers on AI, you'll see things that are even farther from human intelligence than the LLMs of today. There is no "perversion" of the term, it has always been a vague hypey concept. And it was introduced in this way by academia, not business.

[ Parent | on HN ]

Re:Reasoning models don't always say what they think by pixl97 on Thursday April 03, 2025 @03:36PM (#43574352)

While it could possibly be to point out so abruptly, you seem to be the walking talking definition of the AI Effect.

>The "AI effect" refers to the phenomenon where achievements in AI, once considered significant, are re-evaluated or redefined as commonplace once they become integrated into everyday technology, no longer seen as "true AI".

[ Parent | on HN ]

Re:Reasoning models don't always say what they think by Marazan on Thursday April 03, 2025 @02:52PM (#43573805)

We had Markov Chains already. Fancy Markov Chains don't seem like a trillion dollar business or actual intelligence.

[ Parent | on HN ]

Re:Reasoning models don't always say what they think by tsimionescu on Thursday April 03, 2025 @02:56PM (#43573849)

Completely agree. But if Markov chains are AI (and they always were categorized as such), then fancy Markov chains are still AI.

[ Parent | on HN ]

Re:Reasoning models don't always say what they think by svachalek on Thursday April 03, 2025 @03:07PM (#43573972)

An LLM is no more a fancy Markov Chain than you are. The math is well documented, go have a read.

[ Parent | on HN ]

Re:Reasoning models don't always say what they think by jampekka on Thursday April 03, 2025 @03:53PM (#43574561)

About everything can be modelled with large enough Markov Chain, but I'd say stateless autoregressive models like LLMs are a lot easier analyzed as Markov Chains than recurrent systems with very complex internal states like humans.

[ Parent | on HN ]

Re:Reasoning models don't always say what they think by highfrequency on Thursday April 03, 2025 @03:03PM (#43573926)

The results make the method interesting, not the other way around.

[ Parent | on HN ]

Re:Reasoning models don't always say what they think by baq on Thursday April 03, 2025 @03:28PM (#43574267)

Markov chains in meatspace running on 20W of power do quite a good job of actual intelligence

[ Parent | on HN ]

Re:Reasoning models don't always say what they think by phire on Thursday April 03, 2025 @05:19PM (#43575570)

One of the earliest examples of "Artificial Intelligence" was a program that played tic-tac-toe. Much of the early research into AI was just playing more and more complex strategy games until they solved chess and then go.

So LLMs clearly fit inside the computer science definition of "Artificial Intelligence".

It's just that the general public have a significantly different definition "AI" that's strongly influenced by science fiction. And it's really problematic to call LLMs AI under that definition.

[ Parent | on HN ]