125
all 36 comments
sorted by: hot top controversial new old
[-] kbal@fedia.io 44 points 4 months ago* (last edited 4 months ago)

If models are trained on data that it would be a security breach for them to reveal to their users, then the real breach occurred at training.

[-] dgerard@awful.systems 28 points 4 months ago

now you know that and i know that,

[-] Cube6392@beehaw.org 19 points 4 months ago

The big LLMs everyone's talking about and using are just advanced forms of theft

[-] sailor_sega_saturn@awful.systems 32 points 4 months ago* (last edited 4 months ago)

Sloppy LLM programming? Never!

In completely unrelated news I've been staring at this spinner icon for the past five minutes after asking an LLM to output nothing at all:

[-] self@awful.systems 22 points 4 months ago

same energy as “your request could not be processed due to the following error: Success”

[-] earthquake@lemm.ee 19 points 3 months ago

What are the chances that the front end was not programmed to handle the LLM returning an empty string?

[-] sailor_sega_saturn@awful.systems 16 points 3 months ago

Quite likely yeah. There's no way they don't have a timeout on the backend.

[-] dgerard@awful.systems 10 points 3 months ago

boooo Gemini now replies "I'm just a language model, so I can't help you with that."

[-] froztbyte@awful.systems 9 points 3 months ago

"what would a reply with no text look like?" or similar?

[-] dgerard@awful.systems 8 points 3 months ago

what would a reply with no text look like?

nah it just described what an empty reply might look like in a messaging app

they seem to have done quite well at making Gemini do mundane responses

[-] froztbyte@awful.systems 8 points 3 months ago

that's a hilarious response (from it). perfectly understand how it got there, and even more laughable

[-] casmael@lemm.ee 24 points 4 months ago

LLM vendors are incredibly bad ~~at responding to security issues~~

[-] Tar_alcaran@sh.itjust.works 10 points 3 months ago

They're surprisingly skilled at getting money from idiots.

[-] skillissuer@discuss.tchncs.de 7 points 3 months ago

their previous experience in crypto is shining

[-] corbin@awful.systems 20 points 4 months ago

My NSFW reply, including my own experience, is here. However, for this crowd, what I would point out is that this was always part of the mathematics, just like confabulation, and the only surprise should be that the prompt doesn't need to saturate the context in order to approach an invariant distribution. I only have two nickels so far, for this Markov property and for confabulation from PAC learning, but it's ~~completely expected~~ weird that it's happened twice.

[-] motor_spirit@lemmy.world 10 points 4 months ago

Lol that's like expecting gold rushers to be squared away with OSHA, I hope nobody's surprised here

[-] sunzu@kbin.run 9 points 4 months ago

These guys got barely enough staff to run the model lol

[-] 0laura@lemmy.world -5 points 4 months ago

Not really a security issue I'd say. The AI speaking gibberish when you try to make it speak gibberish isn't really that big of an issue.

[-] froztbyte@awful.systems 14 points 4 months ago

sure hope you're not in charge of security anywhere

[-] blakestacey@awful.systems 25 points 4 months ago

Correction: I sure hope they're in charge of security at some place I don't like.

[-] froztbyte@awful.systems 11 points 4 months ago* (last edited 4 months ago)

.......I'll allow it

this post was submitted on 12 Jul 2024
125 points (100.0% liked)

TechTakes

1385 readers
111 users here now

Big brain tech dude got yet another clueless take over at HackerNews etc? Here's the place to vent. Orange site, VC foolishness, all welcome.

This is not debate club. Unless it’s amusing debate.

For actually-good tech, you want our NotAwfulTech community

founded 1 year ago
MODERATORS