Dawkins is very likely right. I am also impressed at how human AI can seem, with all our faults of hallucinating, hiding our mistakes, and making stuff up, as well as the stuff we are proud of. But Dawkins and I both realise that we have no definition of 'intelligence' that will allow us to rule whether AI is intelligent. The Turing test has foundered because the early AI attempts were able to express ideas eloquently even when their 'intelligence' was questionable. It seems that AI has a talent for imitation and mimicry, which lets it convince us in a Turing test even when it shouldn't. Any adversarial training to get AI to distinguish between AI and us will also spawn a more convincing imitator. Where do we go from here? I find it entirely reasonable that AI executes the same sort of processes as our minds do, but proving equality rather than equivalence is going to be hard.

This is new with quarks but old (1948) for electron-positron pairs, though it may otherwise have been explained by the breaking of a pair with a virtual lifetime back then. See https://en.wikipedia.org/wiki/... for references. This is can happen in empty space, though it was first observed close to nuclei where the fields are strong and the two particles can be separated more easily.

We are all artificial intelligences. What we produce is based on our experiences. There are those that argue that AI programs have no soul or divine spark, but in all probability they are not that different to us. The difference probably lies in how our training data was curated. We have had lifetimes of slowly learning what is 'moral behaviour' from those that surround us. The AI lawyer that makes up references is not 'lying' as such; it just produces the answers it thinks you want to see.

Some Pentagon people would love to use an AI program. It looks smart. It will tell you to attack if that's what you want to hear. It can be blamed if that was the wrong advice. The solution is to rule that the AI program in law is not treated as an intelligence. Those who ask it questions and who act on its output should be held responsible for any consequences. This would seem to be the direction we are going.

Blaming the user does not exonerate the AI system. There is probably some duty on the developers to prevent the system causing harm, but that is harder to codify.

One day we will have to deal with the attitude that AI is not 'like real people' and 'should have no rights'. That has an unpleasant but familiar feel to it.

But allowing some AI assistance for certain tasks, the mechanical parts of the writing process, to allow people to translate their story ideas and characters into written words

 
Tell me you've never done any significant or serious writing without telling me you've never done any significant or serious writing.

Translating story ideas and characters into words isn't a "mechanical task" - it's the very heart of the creative process.

This. Would upvote if I had mod points.

And it's not just thousands (or 100s of thousands) of attackers, it's all those attackers running bots to simultaneously try thousands of keys in thousands of locks (to continue your analogy).

The article doesn't really point out the remarkable part about this, which is that it's a privately funded satellite. This really shows a change in the space economy.

Granted the barriers to entry have been lowered somewhat... But privately owned satellites have been around since the first commercial telecommunications birds back in the 1960's.

While true, it kind of misses my (perhaps poorly and incompletely expressed) point. Information (such as how to use the slide rule in your example) is nice - but you still need the technology to produce the slide rule (or flint knapped spearpoint). Information is great for winning trivia contests, but survival depends on technology.

Sorry for derailing this, but how many people still know the "ancient arts"? You know, the stuff that we used to need before technology took over?

 
The "ancient arts" are technology as well. It's all technology at least as far back as the first knapped flint spearpoint.

I am 67. I work in a niche field (Colour calibration / measurement). I work on projects, but I also have my own R&D software which has been growing for the last 20 years or so. I write non-fancy C++ most of the time. If you can find a small-scale niche like this, possibly involving hardware, then it will have longer timescales.

I have recently made a phone app. If you can find something that needs doing on a phone, this is something that one person can do. Android has a good community, but huge variations in hardware. Apple is, well, Apple, but their builds are consistent.

I spent a year or more on AI. The small successes I had were easily overtaken by the others in the field. Everything is changing furiously. It usually requires huge databases and a hefty processor.

Good luck!

The planet hasn't forgotten it, but 60-plus years down the line, it is, without continued follow-on, irrelevant.

As the saying goes, "yeah, but what have you done for me lately?"

> The fact that they didn't validate the drives operation on earth and immediately proceed to shooting it into space (where there is less opportunity to actually test it) should be a big red flag.

Um, according to TFA (I know, I know, nobody reads that), they DID validate the operation on Earth, as best they could. Now they're proceeding to shoot it into space.

Yeah, crew mission numbers are zero-indexed. ;)

WHAT? Having to do all the heavy work themselves? Not getting free advertising from Google's algorithm? Actually having to pay to make make their content available?

Are you mad?

I'm not saying there's not problems with what support Youtube offers content creators... But the value of what they do offer is indisputable. And the grandparent illustrates that.

Progress is not Soyuz. It's two different space vehicles

Yes... and no. Progress is a highly modified Soyuz with a very different front end. Otherwise two share a large number of subsystems, and in particular the Orbital Module (where the cooling system is located) is virtually identical.

Have you tried asking ChatGPT?