Where is the LLM that can reproduce specific whole copyrighted works on demand? All ive seen is reproductions of quotes of a few sentences (fair use) and hacks that can make it ocasionally vomit up random larger fragments of its training data, maybe up to a few paragraphs.

Yep, its definitely not possible that nice small businesses like universal and sony would sue without an actual case in order to try and crush competitors with costs.

Estimates for chatgpt usage per query are on the order of 20-50 Wh, which is about the same as playing a demanding game on a gaming pc for a few minutes. Local models are significantly less.

The purpose of twitter like platforms is to have people to listen to and people to listen to you, so yes vastly lower user counts is a drawback.

Have you considered that maybe other people have different priorities, needs and desires to you, and that for people coming around to your point of view you should encourage them rather than castigate them for taking too long?

hmmm I wonder if that is considered in the thousands of words of this article…

It got more unpleasant after the blue-tick fiasco: identity verification became something you could buy, which destroyed the trust quotient. So I joined the rival platform Mastodon, but fast realised that I would never get 70,000 followers on there like I had on Twitter. It wasn’t that I wanted the attention per se, just that my gang wasn’t varied or noisy enough. There’s something eerie and a bit depressing about a social media feed that doesn’t refresh often enough, like walking into a shopping mall where half the shops have closed down and the rest are all selling the same thing.

Ask your calculator what 1-(1-1e-99) is and see if it never halucinates (confidently gives an incorrect answer) still.

Further to this its not likely to result in less children being abused.

If you have the death penalty for even possessing CSAM as the parent suggests, then there is no incentive to not get into distribution or even actively abusing children and producing CSAM once you posess some. The punishment isnt any worse so why not? Its the same reason for proportionality in other crimes, we dont punish robbing a bank with life in jail as then there would be no reason for anyone who robs a bank to not just murder everyone to leave no witnessess.

What do you use as a vm on an arm mac? I was looking into this a while back to run linux on my work m3 macbook but i couldnt find any good options

That seems a bit of a non-story given that its prefaced on “once the attacker has access to your work email”. Yes once they have that they can do very good spear phishing attacks using copilot, but they could easily do them without copilot too.

They went public, simple as that.

Right, but even taking Musk the mask has only slipped over the last five or so years. Before that he was often treated as a darling for pushing EVs to help fight climate change and SpaceX for reigniting people’s fascination with space. It’s only after he’s got huge amounts of wealth and exposure that its become clear to people what an awful person he is.

This is the logical endpoint for all the people who were complaining that scraping the open web for training is somehow immoral/illegal. Instead of stopping AI those with deep pockets will continue to train on everything while open source and small company efforts will be locked out.

At least in the UK there is now lots of mainstream discussion of “Is it time for people to leave twitter as it’s too much of a cesspool”. Granted they usually then mention threads or bluesky as an alternative not mastodon, but it is definitely possible for social media companies to die out. Once people start to leave in large numbers it can become a mass exodus (see digg and myspace).

They tried that method with the steam machines, it didnt work. A bunch of companies put out half arsed cash in versions and it went nowhere. By putting Valve’s whole weight behind one platform that they tested extensively they got a great product that has made waves. Opening it up now that it has momentum makes sense, but they absolutely made the right call making the steam deck the focus rather than making it hardware agnostic.

No that isnt the case, you could train up a machine to detect if a passage of text was written by Hemmingway (for example) and likely get decent accuracy, that doesnt mean Hemmingway is substandard, just that he has a noticable style (like GPT’s default style). Its likely bullshit for other reasons, but it doesnt imply what you say it does.

Yup, noticable that they use the phrase “99.9% effective”. Effective doesnt have a defined meaning in this contect, unlike accuracy, sensitivity or specificity, so that smells of missleading PR speak to me.

It can likely also be defeated by adding “In the style of X” to a prompt, changing the distribution and pattern of the responses.

Yeah, there is every reason to be sceptical of the hype around AI, in particular from the big tech companies. But to a significant part of the Lemmy userbase saying “AI” is like saying “witch” in 17th century Salem. To the point where people who are otherwise very much left wing and anti-corporate will take pro-IP/corporate copyright maximalist stances just becuase that would be bad for AI.

I honestly cant remember the last time I bought a game and it didnt just work with no tinkering on proton. Though I am on AMD not Nvidia which makes things a lot easier.