It never really explains how anything is achieved - the book does to a greater extent, the movie skips virtually all the "science" and what does exist is basically an aside. But an computer generated voice said "amaze" and apparently that is good so everyone approves.

I seriously struggle to wrap my head around this concept. How is Error Checking going to be achieved at that kind of scale? How will it be cooled - one of the key issues the ISS faces is cooling and a datacentre presumably will be a lot warmer than the ISS - we don't put racks in containers that we then dump in the ocean for nothing. Where we going to site these things? L2?

Hardly new... https://en.wikipedia.org/wiki/...

No. Nope.

We'll have to see, I'm not sure because in my field (strategic healthcare) we can see a use for narrow/deep AI, primarily in radiology but absolutely could have applications elsewhere. But in terms of broader policymaking, its got little to meaningfully add - it knows the topics we are discussing, but it doesn't have any novel solutions for overcoming problems, it does a great job of regurgitating thinking we've already undertaken, but not a single time have any of the models come up with something where we concluded "wow, we hadn't thought of that".

But what I do think is that acquisition of particular reagents, and trying to develop pathogens/chemical compositions that cause harm, would be difficult to do - even a multistep process is likely monitored. What I do note is AI people taking this stance - we see many articles about "Godfather of AI Geoffrey Hinton" (it always says godfather, whatever that means) and his fears. But Hinton has never really articulated anything beyond "I was important at the nascent stage of an industry" and his fears came from that. Perhaps Hinton is a latter-day Faraday, but equally perhaps not. If they are saying these things to "get out of dodge" in the event AI results in Nuremberg Trial's kind of situation - but that is unlikely, and hilariously arrogant for creators of tools that can - and have - on multiple occasions - hallucinated the wrong answer. The brutal truth is that LLM's are cool, they're intriguing, but they aren't "agentic capable". They might be able to do some basic CS stuff in the next 10 years - the blocker there isn't the tech, it's the customer - and if the tool can't grapple that, then the tool isn't good enough.

My hope was that we'd get a load of nuclear build-out before the bubble popped, but I fear I was wrong and the bubble will pop before the foundation concrete has set.

The US may benefit from legions of Chinese engineers maximising the output of Western designs, paying less attention to optimising output from their indigenous chips. The US would prefer Americans (and the West more broadly) work again to maximise their output on Western designs. This is a battle for path dependency I think.

I use aluminium and aluminum pretty interchangeably - I prefer to say aluminum but write aluminium. It's the same thing in reality.

AXA insurance, reporting in!

Oh, so you mean basic evolution. Or does that need tech nomenclature now?

FTL, life-extension, space elevators, replicators? That's hard. 1984? Much more straightforward.

How does it look if we do Humans vs Brachiosaurus though?

Who are these people "walking across the street" for a better job. The market is murder at the moment.

We're just people, people.