Thanks for the explanation.

I love your cow analogy! Here on slashdot we used to have car analogies, but these days the cars drive themselves, so what's the point?

A cow doesn't milk itself! It's perfect!

Cow: An analogy for the next generation.

It doesn't matter, AI superintelligence is an oxymoron. We can't get there from here.

What does matter is if the AI corps get to run roughshod over existing laws "because AI".

What does matter is to ask: if AI is really super intelligent as claimed, why shouldn't the software follow human laws like intelligent humans do?

Does it really make sense to weaken or remove punishments for breaking the law on the grounds that you're super intelligent? I'd like to see super punishments for super intelligent robots when they fuck up, because they should know better than anyone intelligent.

The web already had a solution to that problem: decentralization. The costs of a popular website grow exponentially due to its opposite: centralization, ie the idea that the publisher keeps tight control of the content.

In actual fact, when costs for a popular site rise, the best practice if cost is a concern is to give up control of the content, let people copy the site for free in its entirety and pay their share to host it. When you have two copies of your content around the world, the bandwidth cost is halved, when you have four copies it is quartered. That's how scaling works.

This is usually rejected by businesses who prefer to be the sole content provider of their own content. But if they gave up control, then things would become much cheaper for them. So it's a self inflicted problem.

There are no standards for open source any more. In my view (you can disagree) the community made a choice to splinter when it rejected the GPLv3. What does this have to do with documentation? I view the GPL as the most cohesive and comprehensive philosophy of open source, it gives guidance not just on source code and distribution but also on ancillary aspects like documentation. If more people understood the ideals which originally created the open source movement, they'd understand your point about documentation in Discord and why it's bad.

Nah, there's only Trump. Everyone else just does what he tells them.

Was it porn? It was porn, wasn't it? It's always porn!

I disagree. The early web didn't have the problem of how it was going to be paid for. You simply bought a domain name, fired up your home computer, shared something you owned, ran your own mail server and that was that.

The whole idea of *monetizing* the web, and the later idea that people shouldn't put anything up on the web unless they were getting paid was bonkers, and really a bad idea. It encouraged anyone with nothing to say to feel entitled to create pointless content farms plastered with ads, on the theory that the ads would pay for the labour and hosting costs that only existed to create pointless content farms plastered with ads.... (etc).

It's almost like there are these things called buildings where people do things but they don't tell you, so you have to know exactly where to go and what to say to be allowed to enter and join the fun!

Your point stands, although the days of Prolog are clearly gone (at least for a good while now).

It shows. AI comments tend to have an impersonal quality of waffleness that's a bit grating at certain times of the day. I'm tired though, so YMMV.

The reason was to mildly rebuke you on your off topic word salad in response to a simple apropos reference to the film Wargames, whose main character was played by Matthew Broderick, a still well known actor around the time you joined the internet, and whose other films include Ferris Bueller's Day Off. Both Wargames and Ferris are geek staples for the population who joined slashdot in its infancy, not least because they capture the uneasiness of growing up in a world that doesn't understand you. The Wargames reference is particularly salient here because the linked story is about people who are making the same mistakes being warned bout in that movie, but without the excuse that AI is so new today that no-one has had the time to think about preventable risks.

I hope that helps clear up the reference. Enjoy.

Then don't use it?

That's completely off the point being made. You clearly don't know what the Panopticon is. Look it up on Wikipedia, or better, in your local library. You can ask the friendly librarian lady with glasses about it, she'll brighten up and open your eyes to a world you never knew existed!

The actual point is that the Panopticon is not a technology intended to serve you, nor do you have the ability to opt out of it. It is a technology used by your jailer, to keep you inside the cage without the encumbrance of a traditional correctional facility. The implication here, (to belabour the point unnecessarily), is that you are a prisoner already, and you have no choice but to die or revolt.

That's ok, those on this site with a sufficiently low user id and geeky disposition will recognize the reference easily. You can always ask an AI (or me) for what was meant, I've been around since the old tripod days (a precursor to Facebook, fyi ;-)

Why do you make a distinction between LLMs and other AI technologies? The latter suffer many of the same fundamental flaws as the former. AI technologies are based on ad-hoc curve and surface fitting in high dimensional spaces for decision making. These curves themselves are unscientific, they cannot be trusted on their own. They do not have explanatory power and do not connect with actual scientific theories. They don't even provably converge to the correct choice most of the time.

Trusting AI computed surfaces and simulations is a case of correlation implies causation on steroids. Don't go there.