First, I really loved “Analogical Sock”. I thought that was so clever.
Second, this is quite a profound insight that I’m grateful,for. Fundamentally, token production will inevitably then come down to ownership of the means of production … in which case the most efficient ownership model will be the cheapest and most extractive. So that becomes a much bigger battle, I think, between the economic model driving that ownership of production. Where does profit sit when the national interest is to maximise efficiency? And if profit is separated from the task, because the interest outweighs the need of the individual, what next? It feels to me that in its own way, tech as you describe it now mirrors the crossroads facing our wider economic and political societies as we face huge challenges our existing models seem unable to resolve. Forgive me if this is a little wooly, I’m just thinking out loud. Thanks for the piece.
Sorry to come back .. interesting article on semi analysis (I hope ok to mention) that seems to have bearing on this, around china and access to compute resources. The tech markets going to fracture, isn’t it? I can see Brics going open while developed markets become increasingly isolated.
I think once you consider this then chinas strategic early move to take powerful positions in rare earth etc materials really reinforces the means of production argument. I can’t help but think that the triumph of capitalism that could be seen as AI may also be the agent that forces change to existing economic models - which do need to change given existential impacts like environment, crop yields, wholesale inequality etc. At the risk of going too far I do see all these discussions connecting
Thanks for this Tim. If I can just ask: are you essentially saying that better chips = more efficient corporates = stronger economy = USA wins again?
First, I really loved “Analogical Sock”. I thought that was so clever.
Second, this is quite a profound insight that I’m grateful,for. Fundamentally, token production will inevitably then come down to ownership of the means of production … in which case the most efficient ownership model will be the cheapest and most extractive. So that becomes a much bigger battle, I think, between the economic model driving that ownership of production. Where does profit sit when the national interest is to maximise efficiency? And if profit is separated from the task, because the interest outweighs the need of the individual, what next? It feels to me that in its own way, tech as you describe it now mirrors the crossroads facing our wider economic and political societies as we face huge challenges our existing models seem unable to resolve. Forgive me if this is a little wooly, I’m just thinking out loud. Thanks for the piece.
Sorry to come back .. interesting article on semi analysis (I hope ok to mention) that seems to have bearing on this, around china and access to compute resources. The tech markets going to fracture, isn’t it? I can see Brics going open while developed markets become increasingly isolated.
The challenge with open-weight (which is not open source) is that you can download it yourself, but then you need to pay for your own inference.
If you want to download and tweak the weights, then cool. But many don’t know how to, or can do so efficiently.
Cloud/centralized compute will always be cheaper than local compute.
That’s why I take the economic perspective rather than just tech view.
I think once you consider this then chinas strategic early move to take powerful positions in rare earth etc materials really reinforces the means of production argument. I can’t help but think that the triumph of capitalism that could be seen as AI may also be the agent that forces change to existing economic models - which do need to change given existential impacts like environment, crop yields, wholesale inequality etc. At the risk of going too far I do see all these discussions connecting