> On 14 Jun 2025, at 20:55, Kevin Chadwick via Dng <dng@???> wrote:
>
> 14 Jun 2025 10:01:22 Martin Steigerwald <martin@???>:
>
>> It may not (yet) be, but rather
>> be simulation of one.
>
> It is an impressive plausability of a thought process but it is not (they've added ethics filtering rules due to objections) and can never be as that is not computable. LLMs do not modify their own algorithms and even if they did they would be following a set algorithm to do so.

LLM is nothing more than a stochastic parrot. A fancy glorified predictive text. These algorithms are not “sentient” nor is there any “reasoning” involved above it being a fancy marketing buzzword to bring in more venture capital money.

--
Tom