<![CDATA[When we use words like "introspection", "hallucination", "understand", "discover", and so on when we're talking about LLMs, we make a dangerous mistake.]]>When we use words like "introspection", "hallucination", "understand", "discover", and so on when we're talking about LLMs, we make a dangerous mistake. LLMs have no consciousness, agency, nor self-awareness, and using such terms can make it seem like they do.

(Even "writing code" hits different than "generates code".)

This isn't a pro- or anti-AI comment, it's a truth vs. lying (perhaps to oneself) comment. How we (especially the sellers of trained models) talk about these statistical token generators affects how/when/if we use them and what we expect of them.

]]>https://board.circlewithadot.net/topic/406c8c01-0602-4ba0-8e70-bc5cd038c1be/when-we-use-words-like-introspection-hallucination-understand-discover-and-so-on-when-we-re-talking-about-llms-we-make-a-dangerous-mistake.RSS for NodeThu, 30 Apr 2026 23:57:12 GMTTue, 07 Apr 2026 05:02:36 GMT60<![CDATA[Reply to When we use words like "introspection", "hallucination", "understand", "discover", and so on when we're talking about LLMs, we make a dangerous mistake. on Tue, 07 Apr 2026 18:24:47 GMT]]>@jitterted thank you! this language is so annoying

]]>https://board.circlewithadot.net/post/https://neuromatch.social/users/elduvelle/statuses/116364822956322840https://board.circlewithadot.net/post/https://neuromatch.social/users/elduvelle/statuses/116364822956322840Tue, 07 Apr 2026 18:24:47 GMT<![CDATA[Reply to When we use words like "introspection", "hallucination", "understand", "discover", and so on when we're talking about LLMs, we make a dangerous mistake. on Tue, 07 Apr 2026 06:28:26 GMT]]>@jitterted Agreed. I try and use terms like “generates code”, “statistically likely output” and of course “stochastic parroting” (a remarkably accurate term)

Still struggling to find a phrase that hits home hard enough for the mistakes though - currently I usually say it has generated bad output.

]]>https://board.circlewithadot.net/post/https://agilodon.social/users/thirstybear/statuses/116362006142158923https://board.circlewithadot.net/post/https://agilodon.social/users/thirstybear/statuses/116362006142158923Tue, 07 Apr 2026 06:28:26 GMT