#GlorifiedAutoComplete
ChatGPT -- and practically ALL Ay based on LLM -- are nothing more than #GlorifiedAutoComplete

They 'tokenize' your prompt, and based on the pattern of your 'tokens', emits a series of "most likely" tokens.

A series of #HiddenMarkov models.

They have no actual understanding.
ChatGPT doesn't "claim" anything, and it definitely isn't "factoring in" anything either. It is a system designed to output plausible-sounding text, given text as input. It is not answering questions, much less answering questions with respect to some knowledge base.

>>
September 26, 2025 at 5:36 PM
Previously: Hire humans to write (code, article, whatever)

Now: Hire humans to fix, sometimes completely redo, what Ai botched.

#GlorifiedAutoComplete

futurism.com/companies-hi...
Desperate Companies Now Hiring Humans to Fix What AI Botched
After getting rid of their human workers to invest in AI, companies are now hiring freelancers to fix the sloppy content those AIs put out.
futurism.com
September 26, 2025 at 10:31 AM