It occurs to me that decentralization in a scenario of collapsing business models that drive investments in datacenters, has characteristics of an "antifragile" strategy (as coined by Nassim Taleb). In this case not just for an organization, but for the population at large. With the addition that it would still pay off in an oligopoly of big tech, and not only for economic reasons!
Radical decentralization does require a tolerance for giving up control, which goes hand in hand with a measure of tech-optimism. Or at least the conviction that the odds for positive outcomes in a decentralized world are higher than those in a state or big tech controlled society (to which I subscribe).
There are no risk-free options. Humans often don't respond well to accumulation of personal power. Distributing governance and agency while taking the big AI leap (assuming it is unavoidable in practice anyway) seems the most prudent of the available scenarios. For economic ánd existential reasons.
Gary has a deep-seated neediness to be proven right. He gleefully does victory laps any time the usual ups and downs point to some direction even remotely consistent with what he thinks.
This made Gary quite oblivious to just how big of a deal LLM have been. Now he self-congratulates himself because agents use tools (neurosymbolic, hehe), after spending years mocking this very approach as dead end.
The measured approach here, which Ben points out, is that the cautious optimists are usually right, longer term. This was true for self-driving cars, for neural nets for perception, and now for AI agents.
Giant bets on singularity are doomed to result in disappointment. But skeptics are very wrong. The revolution due to AI is barely starting. A correction is likely but the techniques are very powerful and many companies will do very well.
Gary spends an exhausting amount of time praising himself and claiming prophetic vindication (while mocked by enemies but he triumphs in the end). No matter how far and how fast AI goes, and no matter what direction it takes, he will continue claiming himself prescient since 1990.
My friend, the medical economist JD Kleinke, joked to me, that “an economist is someone who sees something that works in practice and tries to determine whether it will work in theory,“ and, “there are only three types of economists: those who can count, and those who can’t.“ 🤣 yuk yuk. Seriously, though, very cool post.
You have Marcus' argument backward. His causality arrow points the other way, from overoptimistic projections of LLM capabilities that don’t deliver value, to the likelihood of a financial shakeout if sales are even slightly softer than hoped.
He's definitely not saying that financial failure implies inferior technology, and nobody who has worked more than a week in tech could believe such a thing; superior products fail all the time.
“Where Gary’s essay overreaches, it seems to me, is in the implicit inference from financial fragility to technological pessimism — the suggestion, running under the surface of everything Gary writes on these topics, that the coming financial reckoning will also be the refutation of the technology. In fact I believe these are close to orthogonal questions."
It occurs to me that decentralization in a scenario of collapsing business models that drive investments in datacenters, has characteristics of an "antifragile" strategy (as coined by Nassim Taleb). In this case not just for an organization, but for the population at large. With the addition that it would still pay off in an oligopoly of big tech, and not only for economic reasons!
Radical decentralization does require a tolerance for giving up control, which goes hand in hand with a measure of tech-optimism. Or at least the conviction that the odds for positive outcomes in a decentralized world are higher than those in a state or big tech controlled society (to which I subscribe).
There are no risk-free options. Humans often don't respond well to accumulation of personal power. Distributing governance and agency while taking the big AI leap (assuming it is unavoidable in practice anyway) seems the most prudent of the available scenarios. For economic ánd existential reasons.
Gary has a deep-seated neediness to be proven right. He gleefully does victory laps any time the usual ups and downs point to some direction even remotely consistent with what he thinks.
This made Gary quite oblivious to just how big of a deal LLM have been. Now he self-congratulates himself because agents use tools (neurosymbolic, hehe), after spending years mocking this very approach as dead end.
The measured approach here, which Ben points out, is that the cautious optimists are usually right, longer term. This was true for self-driving cars, for neural nets for perception, and now for AI agents.
Giant bets on singularity are doomed to result in disappointment. But skeptics are very wrong. The revolution due to AI is barely starting. A correction is likely but the techniques are very powerful and many companies will do very well.
Gary spends an exhausting amount of time praising himself and claiming prophetic vindication (while mocked by enemies but he triumphs in the end). No matter how far and how fast AI goes, and no matter what direction it takes, he will continue claiming himself prescient since 1990.
Why are there no costs associated with the decentralized AGI? Is the hardware somehow being assumed to be freely given?
https://youtu.be/53fso6tOSKs?is=ahcH7EeVPybWEZ1e
My friend, the medical economist JD Kleinke, joked to me, that “an economist is someone who sees something that works in practice and tries to determine whether it will work in theory,“ and, “there are only three types of economists: those who can count, and those who can’t.“ 🤣 yuk yuk. Seriously, though, very cool post.
You have Marcus' argument backward. His causality arrow points the other way, from overoptimistic projections of LLM capabilities that don’t deliver value, to the likelihood of a financial shakeout if sales are even slightly softer than hoped.
He's definitely not saying that financial failure implies inferior technology, and nobody who has worked more than a week in tech could believe such a thing; superior products fail all the time.
“Where Gary’s essay overreaches, it seems to me, is in the implicit inference from financial fragility to technological pessimism — the suggestion, running under the surface of everything Gary writes on these topics, that the coming financial reckoning will also be the refutation of the technology. In fact I believe these are close to orthogonal questions."