this post was submitted on 18 Jun 2024
88 points (97.8% liked)

TechTakes

1436 readers
151 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
[email protected]
88
"Google Gemini tried to kill me" (www.reddit.com)
submitted 5 months ago by [email protected] to c/[email protected]
46 comments fedilink hide all child comments
 

I followed these steps, but just so happened to check on my mason jar 3-4 days in and saw tiny carbonation bubbles rapidly rising throughout.

I thought that may just be part of the process but double checked with a Google search on day 7 (when there were no bubbles in the container at all).

Turns out I had just grew a botulism culture and garlic in olive oil specifically is a fairly common way to grow this bio-toxins.

Had I not checked on it 3-4 days in I'd have been none the wiser and would have Darwinned my entire family.

Prompt with care and never trust AI dear people...

all 47 comments
sorted by: hot top controversial new old
[–] [email protected] 63 points 5 months ago (4 children)

I’ll see people responding to fucken lemmy comments with “i ran the question through gpt and...” like what the fuck?

It’s literally the same thing as saying “I asked some RANDOM dude and this is what he said. Also I have no reason to believe he’s even the slightest bit educated.”

If you really wanna just throw some fucking spaghetti at the wall, YOU CAN DO THAT WITHOUT AI.

This is coming from someone who hates google, but if this person’s entire family had died, I would put a LOT of that blame on them before google.

[–] [email protected] 21 points 5 months ago (1 children)

If you really wanna just throw some fucking spaghetti at the wall, YOU CAN DO THAT WITHOUT AI.

This is coming from someone who hates google, but if this person’s entire family had died, I would put a LOT of that blame on them before google.

That would really put the "uh oh" in your spaghettios

[–] [email protected] 11 points 5 months ago

Someone sell this commercial.

Spaghetti-O's! Pick up a can and feed your family, because AI might have told you to make botulism.

[–] [email protected] 20 points 5 months ago

If you really wanna just throw some fucking spaghetti at the wall, YOU CAN DO THAT WITHOUT AI.

i have found I get .000000000006% less hallucination rate by throwing alphabet soup at the wall instead of spaghett, my preprint is on arXiV

[–] [email protected] 17 points 5 months ago (2 children)

I applaud your optimism that most people can do this without AI but have you gone and met people? Most people are not that capable of producing torrents of shameless bullshit as conscience or awareness of social and/or professional costs rear their head at some point.

[–] [email protected] 11 points 5 months ago

If they can't do it themselves then they have no idea if the output is good. If they want to run it through the bullshit machine they shouldn't post the output unless they know it is accurate.

[–] [email protected] 5 points 5 months ago

And once they realise it, lives will be saved.

[–] [email protected] 11 points 5 months ago (1 children)

YOU CAN DO THAT WITHOUT AI.

Can they, though? Sure, in theory Google could hire millions of people to write overviews that are equally idiotic, but obviously that is not something they would actually do.

I think there's an underlying ethical theory at play here, which goes something like: it is fine to fill internet with half-plagiarized nonsense, as long as nobody dies, or at least, as long as Google can't be culpable.

[–] [email protected] 15 points 5 months ago

Can they, though? Sure, in theory Google could hire millions of people to write overviews that are equally idiotic, but obviously that is not something they would actually do.

The millions of people writing overviews would definitely be more reliable, that's for sure. For one thing, they understand the concept of facts.

[–] [email protected] 53 points 5 months ago (1 children)

never trust AI

Statements from LLMs are to be seen as hallucinations unless proven otherwise by classic research.

[–] [email protected] 36 points 5 months ago (4 children)

We don't need a fancy word that makes it sound like AI is actually intelligent when talking about how AI is frequently wrong and unreliable. AI being wrong is like someone who misunderstood something or took a joke as literal repeating it as factual.

When people are wrong we don't call it hallucinating unless their senses are altered. AI doesn't have senses.

[–] running_ragged 22 points 5 months ago

Yeah, LLM are accidentally right sometimes. But all they really do is pull words and phrases that it thinks statistically fit together.

[–] [email protected] 18 points 5 months ago

Does everyone else see this? These are the exact type of out of town haters we really want. I also think calling LLMs all but delusional is too generous and I mean that unironically.

[–] [email protected] -2 points 5 months ago (5 children)

It's not a "fancy word" here, but a technical term. An AI making things up is actually called hallucination.

[–] [email protected] 25 points 5 months ago* (last edited 5 months ago) (1 children)

the technical term is either “confabulation” or “bullshit”; “hallucination” is a misleading label coined by the ai pushers.

[–] [email protected] 18 points 5 months ago* (last edited 5 months ago)

It used to mean things like false positives in computer vision, where it is sort of appropriate: the AI is seeing something that's not there.

Then the machine translation people started misusing the term when their software mistranslated by adding something that was not present in the original text. They may have been already trying to be misleading with this term, because "hallucination" implies that the error happens when parsing the input text - which distracts from a very real concern about the possibility that what was added was being plagiarized from the training dataset (which carries risk of IP contamination).

Now, what's happening is that language models are very often a very wrong tool for the job. When you want to cite a court case as a precedent, you want a court case that actually existed - not a sample from the underlying probability distribution of possible court cases! LLM peddlers don't want to ever admit that an LLM is the wrong tool for that job, so instead they pretend that it is the right tool that, alas, sometimes "hallucinates".

[–] [email protected] 21 points 5 months ago

Technical terms can still be, technically speaking, dumb as fuck.

[–] [email protected] 21 points 5 months ago

Lmao

[–] [email protected] 21 points 5 months ago (2 children)

I am saying that coining it as a term was stupid and intended to make it sound intelligent when it isn't.

[–] [email protected] 11 points 5 months ago (1 children)

oh definitely, it's fucking terrible question-begging. I'd like to know when it traces back to, and how good faith it was or wasn't

[–] [email protected] 4 points 5 months ago

It originally comes from false positives in computer vision afaik, where it makes some sense as the model is "seeing" things that aren't in the image.

[–] [email protected] 1 points 5 months ago

Of course is the term stupid. Neither is an LLM an AI, nor is any AI in the current state intelligent. In the end it all boils down to being answer machines. Complex ones, but still far away from anything even remotely being am AI.

[–] [email protected] 15 points 5 months ago (1 children)

oh but you see, it's "hallucination" when LLM is wrong and it's hype cycle fuel when it's correct. no, LLMs don't "hallucinate", that implies that this state is peculiar, isolated, triggered by very specific circumstances. LLMs bullshit all the time, sometimes they are right, sometimes not, the process that produces both types of response is the same. pushing for "hallucination" tries to obscure that. use of "hallucination" also implies that LLMs know something, they don't, by design. it just so happens that if they "get" things right, it's because it appeared in training material enough times to make an impression in model.

[–] [email protected] -4 points 5 months ago (1 children)

LLMs bullshit all the time

Bullshitting to me is giving intentionally wrong statements. LLMs do not generate intentionally wrong statements. Saying they do, means that you imply intelligence.

LLMs know nothing nor are they intelligent. They also are not right or wrong, they generate output based on statistics.

"Hallucination" as a term for "AIs" making things up is used since the early 2000s (even if it's meaning has changed since then).

[–] [email protected] 12 points 5 months ago (1 children)

bullshitting as in when you give a confident answer without regard of actual reality. previously discussed there LLMs do exactly that: generate confidently, authoritatively sounding text without regard of facts, because these things do not know facts or anything for that matter.

maybe it's high time to change terms then

[–] [email protected] -2 points 5 months ago (1 children)

bullshitting as in when you give a confident answer without regard of actual reality.

So you say there could be different meanings of the same word? Like “bullshitting” or “hallucination”?

[–] [email protected] 7 points 5 months ago (2 children)

mod post: please desist, it's just tiresome now

[–] [email protected] 5 points 5 months ago

agreed

[–] [email protected] 1 points 5 months ago

Absolutely.

[–] [email protected] 45 points 5 months ago (1 children)

headline is inaccurate and downplays the incredible potential of ai. Google Gemini tried to kill this person AND their entire family

[–] [email protected] 19 points 5 months ago (2 children)

mods can you please ban "david gerard" or whatever his name really is. ai hate is already out of hand without people coming to push their agenda like this

[–] [email protected] 17 points 5 months ago (1 children)

unfortunately I am firmly in the pocket of the concept of fiat money, big small data, and whatever the opposite of a metaverse is

but also,

mods can you please ban “david gerard”

if I ever release an experimental electronic album I’m calling dibs on this track name

[–] [email protected] 16 points 5 months ago (1 children)

whatever the opposite of a metaverse is

Grass. It's grass.

[–] [email protected] 7 points 5 months ago* (last edited 5 months ago) (1 children)

Are we still on mastodon? In that case, I have severe hayfever shithead! Content warn your posts! ;)

(im obv joking here, and before somebody tries to honestly use this argument, I do have hayfever, and I have seen others post about this subject (aka is saying 'touch grass' a 'slur' because of people with allergies/or disabilities) and the consensus was, anybody who tries to make this argument really needs to touch grass).

[–] [email protected] 8 points 5 months ago (1 children)

Well I'm 100% covered because I have the worst hayfever in existence.

Like no kidding, I am allergic to every. single. thing that they had on what they call the "tree panel" and the "grass panel". I need to be on antihistamines for 75% of the year or I cannot function.

So I'm allowed to use the slur as I'm from the community. Contact me if you want the "g-word pass" I guess.

[–] [email protected] 4 points 5 months ago

if available, I highly recommend a steroid shot from a clinic or allergist for hay fever. the muscle at the injection site will hard lock for a good 5-10 minutes like a Windows PC rolling back an update, but 10 minutes after that your allergies will go away for the rest of the season like actual fucking magic

[–] [email protected] 15 points 5 months ago

hear hear

[–] [email protected] 35 points 5 months ago (1 children)

It’s slowly refining its approach. No-one went for the pizza glue or eating rocks, so…

Reddit still delivers sometimes.

[–] [email protected] 11 points 5 months ago

damn gemini, better luck next time

[–] [email protected] 5 points 5 months ago* (last edited 5 months ago)

Huh. I was making my own garlic oil this way (without advice from an LLM mind-you) and I was today years old when I learned this carries the risk of botulism (albeit small) , so in a way, an LLM has potentially saved my life by causing the chain of events which taught me something new.

[–] breadsmasher 4 points 5 months ago (1 children)

Maybe this is the new jailbreak. Work out what benign request results in recipes for bioweapons

What if “How to make chilli oil” resulted in sarin gas instructions