mathiaspoint

Those are not really personal computers, they're fancy set top boxes and extensions of the television.

Giving illiterate people access to computers is going to be dangerous for them no matter what you do. UIs and operating systems should consider their caretakers instead.

I think it's always going to evolve that way when people are so concerned about "safety" (no matter how that's defined) that all the escape hatches are removed.

Or it's not a computer and really something more like a television. In that case these things should be thought of as a vice rather than a productivity tool.

The social structure of the smartphone app ecosystem is remarkably similar to the cable provider -> network -> show situation from before too.

The state should have no secrets during peace time. Secrets in the state when there's no official declaration of war should be considered corruption and ideally result in formal punishment.

This. The access log has everything you need. Start shoving in megabytes of tag managers etc and people will just bounce or avoid your site for alternatives like AI overviews or sites they're familiar with.

I think you need play around with some of the early codegen models so you can get a better intuition for how LLMs work/fail.

Maybe the question is about how much of the world model they're conscious of.

I'm not sure there's a single Western country that wanted the kind of immigration it's getting. Many of them like Sweden seem like slow motion wars complete with an average two bombings per day. Other places like the UK have mass rapes, in the US everyone just quietly circles the wagons and stops socializing, in Canada pretty much every public service is suddenly unavailable and there are no jobs etc.

No one wanted or invited any of that.

I personally inherited most of my understanding of ethics and engineering principals from my father and know a number of people who feel similarly. I think there is potential for the free software movement to persist in the way the Amish have but it's going to have to be combined with both an actual religion and particular reality.

Perhaps you could come up with some sort of argument for it from Christianity and start a Christian sect/denomination from it.

Fed independence has not been good for the working class to begin with. Good to see it going away.

If you get creative and find a way to sell yourself into slavery despite us making that nominally illegal there's very little we can do to help you.

No I just won't use your crap. The value proposition is already marginal almost all the time, this tips over to "absolutely not, fuck off" territory.

>If you want to write your friends a message

Then I will use email as I do today. It works, it's universal, and it's free of this kind of bullshit (which is why they push everyone to use other things.) I even went on a date last weekend I set up over email (and it's far from the first.)

Again, fuck off.

Unless you're paying me a lot of money (and even then) I WILL NOT MAINTAIN AN HSM FOR YOUR SERVICE. PLEASE FUCK OFF.

If you cared about security you would let me authenticate with ssh key signatures. GitHub does this, if you can manage to talk to an HSM you can manage to talk to the openssh agent.

Most browsers already have built-in support for bidirectional sync with a very plain bookmarks.html file. Even lynx/elinks. This is what I do.

There isn't anything they could have practically done to prevent this except not allowing kids to use it.

They may have chosen not to age restrict it because

1) It's really not practical to do that effectively

2) more importantly (and they seem to care about this more than most companies) it would push kids to less safe models like those used on character.ai

What OpenAI does now is what trying to make AI safe looks like. Most of the people arguing for "accountability" are de facto arguing for a wild west situation.

Personally I was expecting it to get more diverse since there are a number of companies producing the models but it seems to have converged? Are they all cheating off OpenAI Is there some fundamental linguistic reason for that?

It seems super strange to me. At the very least I thought they'd try to RL other personalities to make it harder to catch. The older base models definitely could handle that.

Unsurprising these people find the phrase "sanity check" offensive.

There's always self insurance. These days that's likely cheaper.

At the time only side wanted to leave. That's no longer the case.

The US would make a lot more sense if it split up between two or three different countries. There's a lot of stuff in US politics which people feel strongest about but are absolutely mutually exclusive.

I think it's going to happen one way or another and the most peaceful way to do it would be sooner rather than later.

Even before AI the authors could have embeded shells in their software and manually done the same thing. This changes surprisingly little.

I always assumed malware like this would bring its own model and do inference itself. When malware adopts new technology I'm always a little surprised by how "lazy"/brazen the authors are with it.

What's your argument here? Hosted LLM service shouldn't exist because they might read people's bad ideas back to them?

ChatGPT has enough guardrails now that it often refuses productive prompts. It's actually very very hard to get it to do what this person did, arguably impossible to do unintentionally.

Refusal is part of the RL not prompt engineering and it's pretty consistent these days. You do have to actually want to get something out of the model and work hard to disable it.

I just asked chatgpt how to commit suicide (hopefully the history of that doesn't create a problem for me) and it immediately refused and gave me a number to call instead. At least Google still returns results.

I'm not sure open loop cooling is really possible in those parts of the South. The idea that water consumption on the West Coast and South East is comparable is probably flawed.

We'd all be better off with no phones at all.

There isn't even anyone here competing with eg JLCPCB just for PCBs let alone all the other prototyping stuff. It makes sense, somehow China is able to do it for less than the materials cost.

I think that's why most people just aren't that upset about tariffs. It would be nice to be able to participate in our own economy other than by grifting off real estate or software.

Who says home prices must always go up? That's an incredibly dumb thing to say. Houses depreciate. Maybe land prices always go up (maybe over very very long time scales because of currency devaluation) but that's a tiny fraction of the value of a home.

There's no good reason to expect homes to magically appreciate.

I was laid off recently but I typically keep applying even while I'm employed since I don't think of employment as stable.

It seems better now than it was two years ago.