There are a few places where the logic seems disingenuous to me. You are framing everything negatively, even in cases where it's not warranted. Some examples:
"90% of the value will be burned en route." This means that in aggregate there is a 10% gain, right? Not an aggregate loss. Yet this whole piece is written to suggest that there is an aggregate burning of money.
"If all but one of the dozen major bidders lose, those losses propagate back into the vehicles most people own." But you're saying that in aggregate we expect a gain. So it depends on who the winner is. If the single winner is Google, then "public markets, pensions, and retirement accounts" will all win hugely (because presumably they all own lots of Google). This would be more honestly framed if you included the word "might".
"The prize becomes legible." This is incomplete. If the prize becomes legible, and it turns out that the prize is as great (or greater) than expected, then nothing changes (or investment increases).
The fact that the prize is illegible isn't all that relevant. When the prize is illegible, then it seems rational to me to invest based on the expected value (i.e. your best guess). And half the time, the prize will be equal to or greater than the expected value. Just becoming legible doesn't necessarily mean the investment will slow down.
> "90% of the value will be burned en route." This means that in aggregate there is a 10% gain, right?
The 90% of the value burned is not a statement on the returns of AGI of the winner, rather it's that 90+% of investment capital will see a -100% return (because of the supposed winner-takes-all dynamic).
Even if Google is the winner, evaporating potentially trillions will be bad for the economy.
> "The prize becomes legible." This is incomplete. If the prize becomes legible, and it turns out that the prize is as great (or greater) than expected, then nothing changes (or investment increases).
Prizes becoming legible means that there's a path to AGI that becomes much more fixed instead of endless R&D, which may transform the setup from an auction to a first-past-the-post race for AGI. This would have a very different dynamic as predicted by game theory.
There are a few places where the logic seems disingenuous to me. You are framing everything negatively, even in cases where it's not warranted. Some examples:
"90% of the value will be burned en route." This means that in aggregate there is a 10% gain, right? Not an aggregate loss. Yet this whole piece is written to suggest that there is an aggregate burning of money.
"If all but one of the dozen major bidders lose, those losses propagate back into the vehicles most people own." But you're saying that in aggregate we expect a gain. So it depends on who the winner is. If the single winner is Google, then "public markets, pensions, and retirement accounts" will all win hugely (because presumably they all own lots of Google). This would be more honestly framed if you included the word "might".
"The prize becomes legible." This is incomplete. If the prize becomes legible, and it turns out that the prize is as great (or greater) than expected, then nothing changes (or investment increases).
The fact that the prize is illegible isn't all that relevant. When the prize is illegible, then it seems rational to me to invest based on the expected value (i.e. your best guess). And half the time, the prize will be equal to or greater than the expected value. Just becoming legible doesn't necessarily mean the investment will slow down.
Thank you for reading and commenting!
> "90% of the value will be burned en route." This means that in aggregate there is a 10% gain, right?
The 90% of the value burned is not a statement on the returns of AGI of the winner, rather it's that 90+% of investment capital will see a -100% return (because of the supposed winner-takes-all dynamic).
Even if Google is the winner, evaporating potentially trillions will be bad for the economy.
> "The prize becomes legible." This is incomplete. If the prize becomes legible, and it turns out that the prize is as great (or greater) than expected, then nothing changes (or investment increases).
Prizes becoming legible means that there's a path to AGI that becomes much more fixed instead of endless R&D, which may transform the setup from an auction to a first-past-the-post race for AGI. This would have a very different dynamic as predicted by game theory.