lambdaone

That is an excellent way of looking at things - if a distillation attack against a model is bad, so should be a distillation attack against mankind. I'd argue that this suggests the former should be legal, rather than the latter becoming illegal.

Copyright maximalists would argue that even making a copy in RAM would be copyright infringement.

Perhaps if those logs were streamed to a non-compromised machine and stored there

I'd be interested to hear its opinions about the Tiananmen Square massacre.

I'm glad to confirm your suspicion. I like to have both a walled-garden phone and several open-platform computers. They do two completely different things for me.

To be honest, I'd like my open computers to be even more open and hackable, with a secure boot path that doesn't involve any mysterious unpatchable embedded firmware or auxiliary CPUs, but we are slowly getting there with RISC V.

The very most you can say for Drupal is that it is better than Wordpress at its worst. All the goodwill for Drupal was burnt years ago.

> Close to 3 decades and we are barely scraping 50%

That's what I'd regard as a success.

With majority IPv6 in most leading economies, and getting close to or just over 50% worldwide, the tipping point has been reached and IPv6 is now the dominant IP protocol. The future is a long slow slide of the public use of IPv4 into oblivion, just as envisaged in the original IPv6 planning, and IPv4 support will eventually become a compatibility layer tunnelled over all-IPv6 networks, just as IPv6 was originally tunnelled via IPv4. In the longer-term future, once public IPv4 traffic has reduced to only a couple of percent, ISPs will start to turn it off as an irrelevance, leaving v4 as a legacy protocol for LANs and private networks. Expect this sometime in the late 21st century or early 22nd. Some very old IoT devices may stop working, but everyone who cares about any important use will have replaced them long before.

Meanwhile, the Internet will continue to operate seamlessly for everyone else.

Trivially, all finite neural networks can be represented to arbitrary precision by implementing them with digital logic. All digital logic can be represented by computer programs, so at this point we can pull in the entire apparatus of abstract interpretation, program proving and so on.

Not just that, but we can solve quintics in the almost[1] general case to arbitrary precision, providing we are willing to accept approximate solutions.

[1] I exclude things that are truly perverse such as quintics with coefficients that won't fit on any physically realizable computer.

Just because a problem is NP-complete doesn't mean it's infeasible to produce "good enough" approximate solutions. The traveling salesman problem is a great example of this.

Perhaps a better question would be "can frontier models write a program to effectively stack a dishwasher?"

Perhaps ironically, AI-based searching of the Rust compiler for bugs is likely to find more of these, pushing Rust closer and closer to the point of complete reliability.

The eventual formalization of Rust and a creation of a formally proven compiler is presumably going to be the end point of this process.

You can always choose not to reject again if some number of rejections fail. With a sufficiently large random number range, the chances of rejection at any step can be made tiny, and the possibility of successive rejections drops exponentially with the number of rejections allowed until you have a theoretical probability smaller than the probability of hardware bit errors on the platform itself. If this happens, you can either throw once more and use that result, with a tiny amount of modulo bias, or you can just throw an exception.

It's absolutely how it worked for the Manhattan Project.

The Sentry chip has to be get it right every time; the contained ASI only has to be lucky once.

I can't wait to give this root access to every machine I have.

This is classic AI goalposts-moving.

OK, they can play chess, but that's not real AI - can they write poems? OK, they can write poems, but that's not real AI - can they compose music? OK, they can compose music, but that's not real AI - can they translate languages? OK, they can translate text, but can they do maths? OK, they can do maths, but can they solve a Millenium Prize? <-- we are here

Incredibly creepy. I'm amazed they don't have any idea quite how user-hostile this is, or the reception it would receive.

Creepy as fuck. I'm really suprised they don't understand quite how user-hostile this is.

It can be used for in-jokey comedic effect. For example, referring to M*cr*sft W*nd*ws or Br*dc*m as though they were offensive terms. Or *r*cl*.

The momentum behind rust seems absolutely unstoppable at the moment, in the light of this, the adoption of Rust into the Linux kernel, and the adoption of for formally verified software by Amazon and Microsoft.

While I feel sorry for the people involved, Oracle as a company richly deserves this, for reasons far too familiar to everyone here. This is a come-uppance that has been waiting to happen for decades.

Technically, they could do a whole bunch of things, including doing multihop routing via other connected televisions or insecure WiFi networks. There's no evidence that they do this, as far as I know, but also no reason why they couldn't do this in the future if they wanted to.

"Owning the glass" is a euphemistic way of saying "owning the viewers".

The drive to make warp drives reality seems irresistable. But will they ever actually be possible?

It would seem relatively easy to eliminate this sort of infinite loop by ensuring that all loops are finite, function call depth limited and so on. It isn't a case where you hit the halting problem; the halting problem applies to general programs, not programs deliberately constructed out of parts with finite and calculable limits to runtime - no building the Ackermann function for example, as it contains a recursive loop. With the resources available to Apple, they really have no excuse.

Many cities are doing it already with bike lanes which support both push-bikes and electric bicycles.

With respect to tire pollution, buses are far less bad than cars. Providing a bus is not nearly empty, while the per-vehicle pollution is worse than a car, it's per-passenger pollution is much less.

If only it did. AI is designed to do what you tell it to, and that mostly works. But in the general case, AI does what it does, with no real guarantee of anything.

LLMs seem to have been fine-tuned to use metaphors, presumably because they read like vivid creative writing. The over-use of metaphor in LLM-produced text is exhausting.