Agree, the term is misleading.
Talking about hallucinations lets us talk about undesired output as a completely different thing than desires output, which implies it can be handled somehow.
The problem it the LLM can only ever output bullshit. Often the bullshit is decent and we call it output, and sometimes the bullshit is wrong and we call it hallucination.
But it’s the exact same thing from the LLM. You can’t make it detect it or promise not to make it.
Wow so next time I have a burning need for agentic experiences in my life I know a product exists to serve my need.
We will launch a very deep and very thorough investigation into this matter.
We ask that reporting and inquiries are kept to a minimum as to not hinder the investigation.
It is expected to last for however long it takes for everyone to forget about this, so nobody will bother to look it up and complain that the result of the investigation is “🤷 hard to say who is at fault lol”
It’s baffling how people in the US accept and even adopt the language of NDAs being trade secrets. They aren’t. It’s a weapon to make it harder for people to leave.
TBH he probably knows he is lying, but is making confusing claims in order to push some other agenda.
Probably firing core people to save money while maintaining plausiblish deniability that this won’t do irrepairable damage.
Or just to get himself approval for amassing subordinates for a little kingdom, by displaying an ambitious “plan”.
The expensive autocomplete can’t do this.
AI markering all wants us to believe that spoon technology is this close to space flight. We just need to engrave the spoons better. And gold plate them thicker.
Dude who wrote that doesn’t understand how LLMs work, how Rust works, how C works, and clearly jack shit about programming in general.
Rewriting from one paradigm to another isn’t something you can delegate to a million monkeys shitting into typewriters. The core and time-consuming part of the work itself requires skilled architectural coding.
Yeah what you say makes sense to me. Having it make a “wrong start” in something new is useful, as it gives you a lot of the typical structure, introduces the terminology, maybe something sorta moving that you can see working before messing with it, etc.
This was a very directed experiment at purely LLM written maintainable code.
Writing experiments and proof of concepts, even without skill, will give a different calculation and can make more sense.
Having it write a “starting point” and then take over, also is a different thing that can make more sense. This requires a coder with skill, you can’t skip that.
I’ve been coding for a while. I did an honest eager attempt at making a real functioning thing with all code written by AI. A breakout clone using SDL2 with music.
The game should look good, play good, have cool effects, and be balanced. It should have an attractor screen, scoring, a win state and a lose state.
I also required the code to be maintainable. Meaning I should be able to look at every single line and understand it enough to defend its existence.
I did make it work. And honestly Claude did better than expected. The game ran well and was fun.
But: The process was shit.
I spent 2 days and several hundred dollars to babysit the AI, to get something I could have done in 1 day including learning SDL2.
Everything that turned out well, turned out well because I brought years of skill to the table, and could see when Claude was coding itself into a corner and tell it to break up code in modules, collate globals, remove duplication, pull out abstractions, etc. I had to detect all that and instruct on how to fix it. Until I did it was adding and re-adding bugs because it had made so much shittily structured code it was confusing itself.
TLDR; LLM can write maintainable code if given full constant attention by a skilled coder, at 40% of the coder’s speed.
No way to know for sure based on this. If you used any app that “works with” WhatsApp in any way, you could be affected.
Liar
The bullshit machine is outputting untruths?
The bullshit machine owned, trained and operated by nazis is outputting untruths with a nazi agenda?
Oh no how could this happen.
An economic bubble isn’t when something is fake and ceases to exist when the bubble bursts.
We use the term when the market is inflated way beyond its value.
A bubble bursting is when people realize the market is inflated way beyond its value and pull out, leaving debt-ridden companies to their debt.
Some go bankrupt, stock prices fall, economy news outlets insist nobody could have predicted this, investments go elsewhere, people are laid off, and the people responsible fail upward to another grift.
Wait, you’re surprised it did what you asked of it?
No. Stop making things up to complain about. Or at least leave me out of it.
I tried the image of this real actual road collapse: https://www.tv2.no/nyheter/innenriks/60-mennesker-isolert-etter-veiras/12875776
I told ChatGPT it was fake and asked it to explain why. It assured me I was a special boy asking valid questions and helpfully made up some claims.
The fucking lunacy of the AI bros he lists as examples…
Heyyyy! Who switched the labels on the pods?
laugh track
Now who’s going to clean up Gene?
laugh track
And explain it to his family?
laugh track intensifies
Rob, can you show up to Gene’s wedding pretending to be him while we figure this out?
They try to do security the same way, by adding “pwease dont use dangerous shell commands” to the system prompt.
Security researchers have dubbed it “Prompt Begging”