Hacker Newsnew | past | comments | ask | show | jobs | submit | TeMPOraL's commentslogin

Knowledge has to come from somewhere. Not everyone has perfect parents to couch them on marriage and parenthood.

No. Genes are static fixtures. Furniture. Slow life. Memes took the stage now, they propagate faster.

And guess what? They won’t really need humans to propagate for much longer. Humans are too slow to propagate them anyway.

All the time, yes. But you have to keep two things separate in your thinking:

- Prompted as in prompt made of tokens -> for LLMs, tokens double as a clock signal. Time only flows when tokens are pushed through them.

- Prompted as in specific request placed in the stream of tokens -> Yeah, they do that all the time whether it's getting into infinite loops of repeating same pattern, or suddenly deciding to do things based on inputs they normally ignored.

Also don't forget that everything is a "prompt" for LLM. All input tokens end up in the same place.


So without a token pushed into them they do something? Not sure I understand...

In the current UIs is there a lot of suppression then as I have not seen things start on their own?

I meant an LLM doing something without any external prompt at all. Not doing something different etc but rather do something without a token/prompt ever flowing to it.


Category error. You might as well say "I've never seen a human do anything without being alive, so humans must be very easy to use safely".

Today's AI harnesses just keep feeding AI tokens - and can, in principle, do so indefinitely. "Streaming LLM" is not mainstream but not unknown either. "Automatic context compaction" is somewhat similar in what it does, and very common. "Keeping the lights on" isn't hard. And you have to turn on those lights to have an AI do anything at all.

At the same time: that type of harness often provides a "tool call" interface. Plenty of tools that end up attached to LLMs can do things like run arbitrary code, spawn instances of the same AI, and more.

That's frankly enough. An AI that can spawn more copies of itself can, in principle, just keep itself running indefinitely - even in a non-streaming no compaction harness. An AI that can run arbitrary code on a system can, in principle, do anything a human using that system could do.

The thing that truly limits what an AI can do is the AI itself. Practical AI safety relies on AI being either too weak or too well behaved to cause major issues.


You’d need to train a model to be able to do that.

We shouldn’t be doing that as a civilization in case it works.


Just so it's clear: we are talking about building an AI equivalent of a sensory deprivation tank.

Hard to say, because jobs and processes will adjust to accommodate AI strengths and deficiencies, as AI usage increases.

It's similar to how automation in manufacturing works: we may start with augmenting the human workers with machines improving some parts of the process, but eventually the process itself gets redesigned around the machines.


Because the scientists and mathematicians and sci-fi authors have all already been writing about this for the past 20 years (and I mean non-fiction writing), and nobody cared, giving similar dismissals instead.

I believe Hinton dismissed it a lot over those 20 years

- It's called "writing in bullet points"

- Normies frown upon it


They don't! Before AI, people complain about long emails and what not. The literally preferred to read short ones.

Which is...

Wrong answer. Or at least, obvious and not particularly useful.

Truth is, none of those parties are "nefarious" - they're all just not on your side. And "security" is never an unqualified good thing to have (it's not an unqualified bad thing either). It's just a framework of coercion.

The most important questions to answer about any security system is, what is being protected, for who, and from who. People don't ask that much, not even in the industry - it's an implicit assumption that everyone themselves is a "good person" and is on the protected side of security systems. And then they're confused because it turns out end-users are more often seen as threat actors. All the players mention, but perhaps especially Apple, in its own special way, is protecting the computer from the user just as much as they're protecting the user/user's data from third parties.


Because we're all paying for LLM access for shits and giggles, and not because we're getting actual value from it.

I don't care why you pay for LLM access, it's still spamming my online forums and codebases.

LLMs don't spam on their own. Take it up with people who wield them.

They kinda do though, in that instances have been observed to send unrequited messages even when the person/people in charge of some account didn't expressly ask the models to do so.

For my own use of LLMs, I do try to avoid anything which I know has a risk the artefacts they produce may end up DoSing or spamming, and I've avoided the OpenClaw-type pattern for a broader range of reasons of which this is simply one tiny part, but I'm not absolutely confident I could avoid this even in the code coming out of the free tier of the web chat interfaces except by checking every single line of output every single time.


Nah, it's the technology's fault for enabling it.

At 1000, you can afford better tools and better employees, and replacement parts get cheaper as you order in bulk, and you can explore clever strategies to smooth risk curves.

At 100 000, you can afford a better and continuously improving process, and dedicated facilities, and skilled experts, and parts get even cheaper because you're a volume buyer or perhaps own the supply side, and you get to set your own risk curve.

Lots of things get cheaper at scale. Insurance, too.


Are those ever down?

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: