If you can call being an “also ran” in a field they had a ten year march on their competitors in success, yeah.
Truth be told it was the shoddy code they were forced to use for the vanity of their paymaster might well have held them back, though manifestly that is not a bad thing. Probably the best outcome, really.
In so far as AI can do hardware reliably you can bet your bottom dollar the big chip fabs have already been doing that. They don’t call it AI though and the models aren’t language based, surprisingly enough /s
Seems to me they’re waiting it out. Everything could change in a presidential election and the European economy wins either way. It is an economic bloc after all.
What you describe would be what’s called “cutting off your nose to spite your face”
The concern is not so much that the US will lose friends moreso that other business partners will become more prominent. The US has a lot of social capital to burn. I’m not certain that somebody hasn’t calculated how much they can get away with…
When an engine stalls, the implication is that the chain reaction that drives it is failing - I don’t think that is the case with a GPU as it will quite happily sit there drawing watts til you give it things. In systems nomenclature the inverse term for bubble is utilisation. This or that link is or node is using x% of its capacity. Indeed, if you monitor your GPU with nvidia-smi you will see that very term in the instrumentation.
> I totally struggled to find the right frame of mind to explore any of this stuff without feeling defeated and bamboozled.
I found LM studio to be a nice starting point. Frindlier and more featureful than Ollama and not as intimidating as llama.cpp (though you will want to use that eventually)
Spent a week trying to get sensible results out of llama 3.3 At one point it even simulated doing the work, log output and everything and when I challenged it about the missing artefacts it actually started questioning my intelligence. Seems appropriate for a Zuck enterprise.
Qwen on the other hand got straight to work with astonishing competency on the same system.
From what I read llama3 needs beefier compute to reliably invoke tools, which I presume relates to it focussing more on simulating AGI rather than being a useful tool.
I presume you mean, that what I and others is observing is patterns in mere rhetoric. That this is just unimportant window dressing around the actual problem solving.
Yet, generation of rhetoric seems to be one of the key usecases, and one of the key features that makes this technology seem “intelligent”.
access is provided under condition you respect these restrictions
You are not obliged to honour this. It is not enforceable so it seems strange.
Browsers enforce it, but it can be turned off and nobody expects it to be implemented by a simple REST client application.
It’s a gentleman’s agreement. It’s a statement of expectation to the browser. On the one hand it may be for the protection of the browser user, from cross site attacks, and from malicious code on the web.
But crucially it provides little protection for the endpoints themselves bar accidental misuse.
It is very unusual and rare example of “cooperative” security in a web that’s frequently so adversarial.
It does not. Sometimes it will spawn a mess of ad hoc python, sometimes it will do curl and sed, and very very occasionally it will use the correct tool for the job if it remembers to use the skill you developed in the previous session.
Determinism is a desirable property for software, yes, and its lack thereof from LLM’s is a common complaint, but often a feature depending on who you ask. There is an element of randomness, “hotness” that is central to who LLM work but the pattern we see here manifesting reveals the deterministic processes below, but I don’t think you could rely on this technology to be deterministic, if that’s what you wanted.