Womble @ Womble @lemmy.world

Posts

17

Comments

727

Joined

2 yr. ago

Yup this is the way to think about LLMs, infinite eager interns willing to try anything and never trusting themselves to say "I dont know"

That was exactly the point the church made against the printing press, without needing scribes anyone could come up with whatever foul heresy they liked and publish it for distribution.

The chief difference between now and then is what we consider unpermissable. Otherwise the agrument is the same, we cannot trust people to publish whatever they like or terrible things will happen.

something traveling at mach14 would be heating the air up to around 60000K, given thats about 10 times hotter than the surface of the sun i think it might have been noticed by more than just this obe radar if it was real. For example all the incinerated boats in the Med

the printing press decreased the speed of publication by a larger margin (months to hours for a big book like the bible) and aguably kicked off a century of incredibly bloody warfare with luther and then the counter reformation.

I dont see how being able to get a decent image of Marx with tits from a few mins of generating images is so much more dangerous.

It honestly been really enlightening for me seeing all the same arguments that were made against the printing press and the camera being made against generative AI for text and images. Shows just how little people have changes over hundreds of years.

You're confusing cause and effect, messages get dropped from plans as there isn't demand for them. It's not that people wish they could send crap messages (instead of whatsapp/signal/telegram) but they dont have them on a phone contract.

Its not training the model, it's the model using the context you provide it (in that instance). If you use an unfiltered LLM it will run with anything you say and go from there, for example you could tell it Mexico reclaimed Texas and it would carry on as if that's true. But only until you close it down its not permanently changing the model it is just changing the context in which that instance is running.

The big tech companies are going to huge lengths to filter and censor their LLMs when used by the public both to prevent negative PR and because they dont want people to have unrestricted access to them.

Entirely depends what you are wanting to use it for. Unless you have a beast of a machine you cant run huge generalist models like chatGPT so you have to look for smaller models tuned to your use case. I've been liking mythomax for story telling and wizard coder for coding based tasks.

Eh, not really. Some of the EU laws have serious teeth, there's good reason why pretty much all big tech companies ensure they are GDPR compliant. It doesn't matter how big you are being fined up to 4% of annual turnover is no joke to anyone.

That's only true with the corporate controlled ones, they filter all the results extensively to avoid it giving any answer that goes even slightly against American corporate norms. If you host your own LLM you get entirely unfiltered answers.

If it is they'll be sent down, ignorance of the law isn't a defence.

is that 1/3 monetary ounces or 1/3 fluid monetary ounces?

Really most people thank you for being an annoying pedant who isnt even correct?

You cant say "the worst" when talking about an uncountable group, you have to say "the least good" because I prefer that and it makes me sound smart by correcting you. Apparently that is sufficient for it to be understanding language and for you to be wrong.

I do beg your pardon, it was Georgian not Victorian era when this nonsense was dreamed up for no reason other than preference for trying to cram Latin-esque cases into english.

https://www.thetimes.co.uk/article/no-genuine-rule-dictates-the-use-of-less-or-fewer-cs25kv8s5

The very notion of a neat distinction between fewer and less according to whether the noun is countable or not is a myth. It was invented out of whole cloth by an ill- informed 18th-century pedant called Robert Baker in his book Reflections on the English Language (1770). He proposed this distinction not as a hard-and-fast rule of grammar, moreover, but as a tentative suggestion with caveats (“I should think . . . it appears to me . . . ”) that you won’t find in modern style guides.

The wiki article on it notes that

The Cambridge Guide to English Usage notes that the "pressure to substitute fewer for less seems to have developed out of all proportion to the ambiguity it may provide in noun phrases like less promising results". It describes conformance with this pressure as a shibboleth and the choice "between the more formal fewer and the more spontaneous less" as a stylistic choice.

i.e. it is a shibboleth for saying "I am educated unlike you uncultured lot who use natural sounding language"

The less/fewer distinction is arbitrary Victorian bullshit flying directly in the face of how English is used. The only point of it was to try and make English more like Latin and allow aristocrats who spoke Latin to look down on those without expensive private education.

Please dont perpetuate it.

yup, just like the ai needed lots of pictures of astronaughts on horses to make pictures of those...

No you dont? You can literally just copy files onto an external hard drive, there is no requirement to use steam to do that.

i mean, if theres a plate on the floor and soneone checks theres only one person on it before a button is pressed that solves that and theres no reason to display the number. Flash up an error to retry if its more than 3 sigma away from the expected mean.