Looking for list/website that documents llm fails

[device_enchanter@feddit.org]

I'm having an all day meeting with some of our companies 2nd tier execs, and they are firmly on the ai hype train.

For web3 (crypto/nft) https://www.web3isgoinggreat.com/ is just a wonderful resource to show what can and does go wrong.

Is there anything of the sort for llms (or what some call AI) ?
Up until now the use cases where benign stupidly burning money operations. Now we enter territory that could harm people and I want the best chance to deflate their AI delusion.

Thanks!

[paw@feddit.org]

Pivot to AI is similar to web3 is going great. You could also look into Ed Zitron's articles: https://www.wheresyoured.at/ however they are pretty rant-like but do also contain a lot of further links.

[mesamunefire@lemmy.world]

That sounds like a hilarious idea. GL!

[biptoot@lemmy.today]

Aiincidentdatabase

Also search GitHub, several repos list failures.

What kind of company? I do this training for our public sector / state agency execs, and have a fairly well stocked slide deck currently I might be convinced to share

[aboubenadhem@lemmy.world]

Ask an LLM to find you a list—if it doesn’t, then you have a failure right there.

[j4k3@lemmy.world]

In nearly every instance you will be citing stupidity in implementation. The limitations of generative AI in the present are related to access and scope along with the peripherals required to use them effectively. We are in a phase like the early microprocessor. By itself, a Z80 or 6502 was never a replacement for a PDP-11. It took many such processors and peripheral circuit blocks to make truly useful systems back in that era. The thing is, these microprocessors were Turing complete. It is possible to build them into anything if enough peripheral hardware is added and there is no limit on how many microprocessors are used.

Generative AI is fundamentally useful in a similar very narrow scope. The argument should be limited to the size and complexity required to access the needed utility and agentic systems along with the expertise and the exposure of internal IP to the most invasive and capable of potential competitors. If you are not running your own hardware infrastructure, assume everything shared is being archived with every unimaginable inference applied and tuned over time on the body of shared information. How well can anyone trust the biggest VC vampires in control of cloud AI.