"King of the Cannibals": How Sam Altman Took over Silicon Valley

Actually-open AI with open weights and open source is innovating rapidly too, especially when it comes to making models more efficient on smaller hardware. We have fast local mixtures of experts now. Rumor is that Llama3 is training as well, and that it may be some kind of MoE.

I feel like focusing on efficiency first for local models rather than shooting to beat GPT-4 is the right approach, because without major efficiency improvements a shot for simple dominance will just yield a model only wealthier people can afford the hardware to run.

OpenAI still seems in the lead but it’s shrinking.

A better way of looking at it would be to try analyzing the performance of Sam’s personal investments. And comparing it against a baseline (e.g. Sequoia or YC).

Sounds good on paper, but the nature of modern day life is far different than it was when other revolutions occurred.

Yes, revolutions in tech are just that - they up-end the old way of life, the dust settles and a new way of living emerges.

This is so basically obvious that its not worth mentioning.

What is worth more than mentioning, but truly paying attention to, aside from Thomas' analogy to a neighboring mcdonalds, which is an actually insanely lazy look at this matter - especially coming from one who is a security expert, is that the level of "entanglement" into our daily life with massive core-AI systems, such as the capabilities of GPT-'N' has, cannot be understated.

When cars overtook horses, we could weaponize cars by putting guns on them and calling them tanks. But once you put a gun on a car and call it a tank, it can only shoot at one target at a time.

With AI, when it provides the substrate upon which you will interface with the information relationship you have the rest of global reality, can be easily weaponized, not just against you, but anyone and anything interfacing with it, its important to know who, or at least what incentives, truly owns, or at least is the lens through which one interfaces with whatever information that system provides.

SO, aside from 'doom-and-gloom' -- its really important to really understand as much as possible, and rather than Thomas' just throwing up his hands (which again is bizarre from a security guy just saying he doesnt care who runs the black box, is weird) - I want to know how whatever that system is (openAI/whatever dominant AI) and who made it such and what agency I still retain in my digital life.

--

@Unity:

Entanglement of big business and government in our lives was and will always be - what big-core-AI provides both business and governments (which now there is vanishingly thin membranes between the two these days) complete solvent for any friction for said entanglement.

AI's capabilities for extracting indivulaized-insights-AT-SCALE is what AI provides to both, with zero recourse from the individuals. That is the "alignment" problem that Aza and Tristan are pushing back against.

Also- I do not "worship" or "idolize" SA, and especially Thiel - I am terrified of them. That's why I said that they may attempt to appear as a Carnegie, but are actually a Rockefeller (Carnegie attempted to wash his reputation with endowments - Rockefeller was just an evil oil-carpet-bagging C*Sucker and an evil person (he is the reason we have "fossil fuels" as manufactured scarcity for profit)

I have no idea why you think I idolize any of this - I am fully, and even further, into the Raskin-Harris camp... alignment regulation is going to get trampled - and to be honest, I have serious questions about SAs motives...

I'd bet he's already got a team working on the plans for his bunker/lair, and taking notes from the Zucks and Thiels on best design resources.