I don’t presume Anthropic is designing anything to be ethical. First, the chatbot is unpredictable by design. Randomness is built in. Second, if they didn’t intend for bad things to happen, why didn’t an employee babysit it?

As the article said you can replace rogue agents with unpredictable computer programs if you really want, I don’t see the necessity.

Speaking of goalpost shifts: “Rogue agent” is the clickbait title across multiple articles. Remember you said nobody was misrepresenting the AIs. The still-inaccurate clarification (the unsupervised chatbot functioned exactly as it was designed to function, and the script running it allowed it to execute exactly the commands Anthropic wanted) doesn’t help the article the clarification is from. It certainly doesn’t absolve every other author.

the first article doesn’t say anything not factual

The title alone is not factual. AI did not go rogue. It behaved exactly as designed.

Nobody is claiming any of the things this article is fighting against

Three counterexamples:

• OpenAI models went rogue. The Guardian
• Why the Hugging Face Hack Should Make You Worry More About A.I. NY Times
• Hugging Face hack marks start of dangerous AI cyber era and many firms ‘don’t even know it’ CNBC

I realize we are in a bit of a bubble here, but this is the framing in the wider world

there were unintended negative consequences

Well…

“Aggressive” is how I understand the language of someone who attacks “natural intelligence” and promises to “very loudly ignore” the people he deems "social warrior"s and “religious” for not agreeing with his opinions, telling anyone who disagrees with him to “look in the mirror”.

People trusted Linux to be better than Microsoft Windows, and saw the effect of a company with tons of money and first-class AI access exponentially worsening their product.

It has been disappointing to see people like Linus Torvalds turn aggressive against people who do not agree with his ideology, and I find it distasteful for you to use reductive attacks against strawmen as well.

Considering Graphene’s entire reputation is staked on extremely precise definitions of security, this is a little concerning.

They’re the “best” (read: most secure) mobile operating system. If the best degrades, I don’t foresee some Graphene alternative stepping up very quickly (see also: cross-platform alternatives to Google-made Chromium and Google-funded Gecko)

I’ve heard similar but am not sure. You’d be better by confirming with offline version of the available models.

I believe I did this with DeepSeek’s distillation of Qwen, and it basically printed out state propaganda when I asked it about Tianmen Square… But that meant Qwen (from China’s Alibaba) was intentionally censored, not Deepseek itself. I can’t afford the hardware to test their model, especially not now.

while Anthropic was fast to call the cops on a frustrated Claude user, the Standard also reported a key detail: Anthropic refused to show police the actual messages, citing Anthropic’s internal policy. In short, Anthropic reported a user for in-platform speech, and then refused to provide police with evidence of actual wrongdoing.

Anthropic’s apparent game plan:

1. Get a list of anti-AI advocates
2. Swat them
3. Refuse to show evidence of claims

This announcement will make YouTuber Markiplier $24,225,000 richer.

As part of the deal, the company’s shareholders will receive a $285 million cash payment, or $1.14 per share, and the stock will remain listed on the Nasdaq.

One of the shareholders benefiting most from the news is Youtuber Markiplier, who acquired a massive 8.5% stake, according to a July 13 filing. Private equity firm BlackRock also disclosed a 6.4% stake earlier this summer.

FYI: You are arguing with what the CSAM ad says. That’s what the article is quoting here.

I never had a problem with older TVs’ boot times. In fact, I remember them being faster, the farther back I go…

And it’ll tell me this directly, rather than passing that data to my insurance company to deny my claims, and telling banks to deny my long term loans because I have a known expiration date, right?

A full minute of processing power and the best it could come up with was crashing through a divider and ending up in the far lane

Have you considered using it for hosting a small personal blog instead?

(There might be better options, but this is the first one I became aware of)

I have found Bluesky to be less toxic than Twitter, but that’s not exactly a huge improvement.

I remember that on Twitter, if somebody decided to reply to you and to be a jerk, not only would you get a notification of that reply, you would get a notification if people liked that reply. Bluesky by default allows you to only view the feeds of people you follow or are mutuals with. These are all good things.

But since there is VC funding, and because they are looking for ways to expand, at any moment that rug can basically be pulled out from under people’s feet. It’s technically decentralized, but in practice, it’s probably 90%+ controlled by one company.

There are many echo chambers, but this one is mine.

I disagree! /s

Walking down the aisles of any store selling computers, I can attest to this. If a company can’t sell a Windows computer, they will simply sell a Chromebook instead.

Two thoughts based on what you wrote:

There’s a bit of an issue with that. Re-Centralization. If the grand majority of users are in the “Main” or “Primary” instance, it’s just moving one mass from one to another, without spread…

I think it’s plenty fine to recommend a specific instance to people. It doesn’t need to be the main one either, it can even be yours:

Does the Fediverse have an equivalent to matrix.to?

Not really. Probably because Matrix can “get away with” recommending some default software (including web apps) to view a community from any server, regardless of where it is, mostly because there’s very few instances of that software to recommend. Matrix.to is a first-class service provided by the same people who run the flagship Element client and develop the server too, after all. And it’s still limited: matrix.to will not point you towards a niche instance of Element, for example.

But things get worse when it comes to Mastodon and Lemmy because usually instances tend to act as both data servers and clients. If somebody sends their profile link to a friend, it goes to their server. People have implemented sharing techniques, but these are janky at best, and require extra configuration and pre-knowledge on the part of the recipient.

And I still don’t know how to open a post from a different Lemmy instance on my current one without manually looking for the community and then the post.