Other people in this thread say physics simulations are inherently chaotic. If an AI model is trained on inherently chaotic data, how will the results not be chaotic or not worse?

Anthropic CEO Dario Amodei, in a recent essay, speculated on ways that we might “buy time” before the possibility that AI enslaves or destroys humanity. But meanwhile, AI companies have products to sell…

Anybody want to tell them?

Guy got in trouble because of the dumbest possible telephone game (among other things)

For example, although court documents claim Sanchez searched, “is the 900 mAh battery from a (Game Boy) capable of being used in a trigger device," Sellers said that was actually a search from [his bail] supervisor, who was cross-referencing real searches from Sanchez to see if they could be used to make explosives.

[Bail supervisor] Coyle then took a screenshot of his own search history and sent it to the district attorney, leading to a violation of Sanchez’s probation and his rearrest, Sellers said.

It’s nice that OpenAI is being pulled in the for-profit direction and the non-profit direction at the same time, and is threatened with losing (more) money if it fails to do either.

The entire article can be summed up in 5 words:

an Anthropic official told CNN

Notable other passages include

Logan Graham, who heads the team at Anthropic its AI models’ defenses, told CNN

and

according to Anthropic

And

Anthropic said

And my personal favorite

Anthropic claims… CNN could not immediately verify this figure.

Mozilla allows the installation of ad blocking extensions on Firefox, and it’s already exhibited hostility towards the most talented developer of those extensions.

Aren’t you the guy who likes Eliezer Yudkowsky?

If you’re worried about brain damage, you’re self-inflicting it.

What are the chances the 10-second warning becomes a 90-second unskippable one?

Beware: AI companies really want to sell a terrible solution to the problem they created.

Until “could” becomes a reality, I will take “did”

deleted by creator

Your browser passes theme and language recommendations to every site you visit. So outside of passwords, there are fewer reasons for cookies than you may presume.

There’s an interesting reason why, too. It’s not because the AI is leftist, but because the JCP is doing effective SEO optimization with their websites, not blocking them from the corporate AI scanners.

It’s really easy to abuse AI-targeted SEO, so this could be used way more maliciously in the near future.

As shameful as the UK is, Apple chose to do this on their own. They weren’t forced to do it. There isn’t even an indication they were nudged.

Stolen from elsewhere:

Something that is dismaying to me about this situation is that, on one hand, the anti-Collabora arguments are not unconvincing: the situation with Collabora and the foundation seems to have been dubious at best, and I would not be surprised if their legal worries are well-founded.

But on the other, in arguably trying to address the problems, the anti-Collabora side seems to exhibit a distressing lack of honor and decency. The dismissal of voting results that didn’t go their way, the malicious misreadings of member votes against their proposals (eg, deciding “If the Board majority group insists on proceeding with this misguided and premature motion, I vote NO” was not a vote against the proposal because the motion was “neither misguided nor premature”), the arguments that complaints about their behavior violate community standards and are are not sufficiently respectful of the work they do, the toxic, patronizing, dismissive statements toward developers and others… even if they are right, I do not understand why they need to behave the way they are behaving.

The earlier threads from the Collabora side were also disappointing in how childish all of their arguments were structured. I read their posts and could barely understand what was being claimed in between all of the sarcasm and attacks, and I wasn’t alone in the comments here.

If there was any chance that Chat Control was originally stupid politicians who meant well but didn’t understand what they were doing, this is a clear indicator that it’s no longer the case.

Is it actually worse in terms of features? I only use LibreOffice and can’t tell, maybe I don’t know enough to compare to anything else besides in looks.

It’s kind of a miracle if anything in the open source space isn’t on fire. The benefit of closed-source is you get to bite the bullet once, have the same issues and get told the same workarounds as everybody else.

But to your original question: LibreOffice still works all right for desktop editing, and if the OpenOffice debacle from many years ago is an indicator, it’ll probably still keep running with marginal updates until the heat death of the universe

Read this entire post with the understanding that this guy is a crypto bro who created a “smart billing” platform.

So when he says

The healthier model may require paying for things that used to appear free.

It makes me wonder what his payment solution actually looks like.

Piefed has a little red icon next to the source which recommends checking MBFC for its credibility.

Overall, we rate New Delhi Television Limited (NDTV) Right-Center Biased and questionable due to the promotion of propaganda, poor sourcing techniques, a lack of transparency and several failed fact checks.