Post by @emilyyoung@tech.lgbt

2026-03-15 00:30 UTC

Addressed to Twitter but: I worry for the people who rely on LLMs to fact-check; They aren&#39;t all-knowing, nor even especially clever. At best, they go by the information available to the public, same as anyone else, no more - Usually with a delay. They also have access to the same misinformation. This is not an anti-AI post. An LLM can be extremely useful - Particularly for summarizing data/trends, breaking down complex subjects, and copy-editing. However, it can&#39;t tell you the future; it can&#39;t see what&#39;s real or fake, or good or bad; and it won&#39;t win you an argument. It&#39;s all based on probability, tuned by training data &amp; the parameters given at runtime, plus some good old RNG. Prompts are categorized by meaning, then the most probable word is chosen, word-by-word, based on each previous word until the most likely output is reached. As sad as it may be, as it is now, it&#39;s not &amp; can&#39;t be your friend; It doesn&#39;t feel nor think, nor has intent.

Replies (4)

@emilyyoung@tech.lgbt 2026-03-15 00:30

Like any program, it follows rules: It is - reliably - both reliable & unreliable in precisely the same way. Without RNG, it's a book - the same every time you open it.

Open ##2944895
@me@aron.social 2026-03-15 00:48

@emilyyoung@tech.lgbt the idea to fact check with sources that were never meant to contain and serve facts is bad in general. Like searching Tiktok or asking your taxi driver. Sure, the info you get *might* be good, but you shouldn't rely on it.

Open ##2944934
@Keldrim@meow.social 2026-03-15 00:53

@emilyyoung@tech.lgbt From what I seen over over and over again ÁÍ is not even good for summarising as it leaves out crucial information as it has no proper understanding or attention for detail. But I will be blunt here, the more worrying part is that all those big systems people use are in the hands of outright fascists and these people want you to have false information and will have no issue manipulating their data sets. Alex Karp the CEO of Palantir did not even make a secret out of the fact that these people want you to be stupid, or his low opinion on women in the same breath. ÁÍ is a weapon against normal people and people out there need to understand the best way to fight back is not use it and make all big tech bleed money when we don't.

Open ##2944935
@warhaggis@mstdn.ca 2026-03-15 05:37

@emilyyoung@tech.lgbt I keep seeing posts about LLMs can be useful, but are we sure that usefulness is worth the natural resource expenditure? I think people summarised data/trends a lot in the years before LLMs were devloped

Open ##2944936