Pavel @spavel.bsky.social · 22/03/2025LLMs are the Wish spell of serious and highly paid tech professionals. 528745
Pavel @spavel.bsky.social · 22/03/2025This is why every AI fail will be met with "what was the prompt?" This is why devs think that asking their LLM to not hallucinate will prevent it from hallucinating. The prompt is a wish, and the perfect wish can do anything. Except, of course, within the bounds of the spell's power.arstechnica.com“Do not hallucinate”: Testers find prompts meant to keep Apple Intelligence on the railsLong lists of instructions show how Apple is trying to navigate AI pitfalls. 1031353
Charby @stornahal.bsky.social · 22/03/2025LLMs are doomed to fail until they’re allowed to build their own internal permanent model of the universe - and given ways to double check against reality. 010
it's_my_potty🧤 @oldbaby.bsky.social · 25/05/2025computer write me a blockbuster summer hit movie with thrills and chills that will break all the box office records and sweep the oscars 220
(¯`·._{JEFF}_.·´¯) @handle.invalid · 22/03/2025"Do not hallucinate" Sort of like asking a human not to breathe. 020
Dylan Wilbanks is not a Halloween person @dylanw.social · 22/03/2025I just think of all the poor monkeys the AI lords have killed to get just a few more stupid wishes from their paws because THIS TIME they are going to get the prompt right 020
Pavel @spavel.bsky.social · 22/03/2025We are constantly told that AI is a revolutionary technology, just as the rulebook explicitly says Wish is the most powerful spell there is. It's not true in either case, but the story around it makes it seem so. Makes it seem like it can solve any problem - unless you understand how it works. 120722
🦇 Murphy @mjosmurphy.bsky.social · 03/03/2026"do not hallucinate", i write in binaric on the scroll, affixing it to the Imperial Titan with sealing wax and the blood of fallen battle-brothers. surely, this will quell the hulking machine's angry spirit 010
FavoriteMovieAvatar2 @shoe3482.bsky.social · 23/03/2025The "no true scotsman" defense of AI. AI cannot fail, it can only be failed by poor prompt writing (or poor training data, etc). They'll find any excuse to avoid admitting that the tools don't work, won't ever work in order to fulfil the promise they've made to investors. 062
