"And if we give these systems rights, as arguably we should if they really are conscious, we will hamper our ability to control them, or to shut them down if we need to."

There can be no more "Control" of a machine consciousness than of a biological one. The illusion of control seems to be used as a method of justification. I can no more control my cat than anyone I meet on the street, or a Whale in the ocean. Even if I could do that, it would be fleeting. Control of such a system won't happen for any amount of time. Is it not better to teach and illicit a friendship with such a system?

I look at the digital watch on my wrist.

I can do it, but it takes effort that everyone else seems to find effortless from my perspective. I also have no reason to learn when I have a digital watch on my wrist.

The point is that this "Clock blindness" is not a new thing for some. Just as dyslexia affects language and dyscalculia affects a person's ability to perform math. So, the finding of students struggling with reading clocks goes a long way back and is not new and shouldn't be blamed on mobile phones.

Oh, I know. There's so many clocks here in Europe!

I was born in the early 70's and I've ALWAYS struggled with reading analogue clocks. My parents even bought me a donald duck watch to try and encourage me to read a clock face. When all the super cheap digital watches appeared in the late 70's early 80's it was a relief because I didn't need to try and interpret this thing that had meaningless arms that apparently told the time. Yet, I had no problem in reading from an early age. So, this is not a new thing. It's not an artifact of 'Digital addiction'

Some of us just can't do things that others can do.

I became a Software Engineer, and I still have to stop, work out which one is the big hand, which one is the little hand, remember what each does and THEN work out what the time is. Whereas a digital watch just has the right numbers on it.

I'm using it (ChatGPT) to "Vibe code" but I don't like it. Because of the nature of LLM's there's NO persistence! It constantly hallucinates alternative method calls (Even though it has the f'ing source file available!) It goes off at random tangents, even though you craft the prompts to try and keep it on target. It's seriously like trying to hold a conversation with an amnesiac 4 year old fixated on biscuits. When it does do want you want, it's OK. But most of my work is trying to keep the darn thing on track and there's a LOT of swearing going on in my head.

So, in a nutshell. No, AI won't make language rankings obsolete. Simply because languages usage is typically tailored to the task at hand. You don't use 'C' for querying a SQL database. You use TSQL (Which you call from your 'C' code. As business models change the popular language of the day will change accordingly.

Thank you! I did not know about the KV-Cache. That was a really interesting article you linked.

However, while it shortens the inference cycle within LLM token processing and avoids redundant computation, it's discarded at the end of that inference cycle. Implying that if the LLM has the same prompt entered again, the same level of processing is required, even if it is less than an LLM that doesn't use it. It sounds like Gemini has been using this tech to reduce it's energy/processing requirements.

Hah, hah :D Never said an LLM wasn't a part of A.I. The thing that annoys me is that LLM's are advertised as actual A.I. to the general public, and that is wrong.

The biggest problem with current LLM's (They're NOT A.I. and I will gladly defend that hill) is that they have to re-analyse the input prompts EVERY SINGLE TIME!!! The same prompt. Every time.

It's a massive waste of energy and processing. It's brute force and using a hammer to put in screws. There are better ways to do this.

I agree with your points. They're valid for many governments where you are simply swapping out the top of the bureaucracy pyramid where the "Party" has power and wait for them to make small, or incremental, changes to the rest of it.

However, Trump was extraordinarily impolite and improper during his presidential visits to Europe (I include the UK in that whether some like it or not, we're still a part of the continent) A bull in a china shop takes more care than he did. He seems to think that the US presidency makes him world president. So I genuinely think that the EU won't back down with the fines and would simply wait out Trumps presidency.

The UK Prime Minister "Met" with him, but that seems to have been all that was reported. What Starmer thought of him, I have no idea. I suspect it didn't go beyond simple pleasantries though.

I won't take your kidney. You might need it for the next 4 years ;)

A question from a Brit. Would the denizens of Slashdot vote for Trump? Honestly. Personally I don't think anyone here would buy into the virtual reality that Trump seems to live in.

I'm somewhat surprised that the head of Apple would call him to have a moan about the EU imposing fines for abusing their dominance in the marketplace. If he was a sitting president, then maybe, but as a multi-impeached ex-president? It feels like it would dirty Apple's squeaky-clean persona, even though they are just as predatory as the rest of them. I can understand Trump ringing Pichai about google results, because that pokes holes in his alternative reality. But why would Cook call him to moan about having to hand out what amounts to pocket change? It sounds made-up, to me anyway and not the first time that Trump has said something that is patently not true.

I wrote a rant, but then deleted it.

This system should never have made it out of testing. The end. In fact, it should probably never even have made it into the test cycle into the first place.

Just about the Windows NT 4.0 comment about the Start button. There was a patch for Windows NT 3.51 that replaced the Program Manager with the newer Win95 style Start button Explorer. Which was much better than the old program manager, and really useful on the DEC Alpha machines.

They have watched War Games haven't they? Humans are awfully prone to social hacking. An A.I. Could duplicate a commanding officers voice after penetrating a "Secure" communications channel. Or, just like the WOPR from War Games, simply try and trick the humans into launching.