Category: Economics, Finance, and Political Economy

SC:
I just don’t see it.

We were at the end of extraction of opportunity from not only the internet, but from smart phones. In fact, we’ve been misdirecting capital away from reforming the 1980s technology that has imprisoned us in the network effect of the desktop-laptop and file systems. Even the browser tech is perpetuating terrible software, and much costlier development that’s only a trade off vs cost of distribution and installation.

So far the prmary value of openai is in increasing the productivity of spam production – although code improvement is certainly helpful.

Yes we are in the initial stages, but you can’t, we can’t, extrapolate a curve.

If your startup biz is threatened by openai, it was never a good business anyway. (And really, most startups aren’t good businesses, and they’ll crash soon under the current correction, just as they bad ones did in 2001.)

That said, we are fascinated by the rapid progress of what is a simple technology made possible by advancements in processing power. But having been through a sequence of AI winters in my life, and while I see promise arising, the fact that we can produce the equivalent of slightly better search results, but with no logical consistency, is more of a novelty that will result in increases in spam, marketing, and other tribvial content competing for our increasingly scarce and shallow attention.

And the fact that they’re teaching it to lie under cover of ‘alignment’ only means that the research into telling lies is progressing faster than research into telling the truth with some degree of logical consistency.

Cheers.
Sorry to counter-signal but I can’t find a use for openai or any other group’s work so far, other than code completion.

Reply addressees: @blader

SC:
I just don’t see it.

We were at the end of extraction of opportunity from not only the internet, but from smart phones. In fact, we’ve been misdirecting capital away from reforming the 1980s technology that has imprisoned us in the network effect of the desktop-laptop and file systems. Even the browser tech is perpetuating terrible software, and much costlier development that’s only a trade off vs cost of distribution and installation.

So far the prmary value of openai is in increasing the productivity of spam production – although code improvement is certainly helpful.

Yes we are in the initial stages, but you can’t, we can’t, extrapolate a curve.

If your startup biz is threatened by openai, it was never a good business anyway. (And really, most startups aren’t good businesses, and they’ll crash soon under the current correction, just as they bad ones did in 2001.)

That said, we are fascinated by the rapid progress of what is a simple technology made possible by advancements in processing power. But having been through a sequence of AI winters in my life, and while I see promise arising, the fact that we can produce the equivalent of slightly better search results, but with no logical consistency, is more of a novelty that will result in increases in spam, marketing, and other tribvial content competing for our increasingly scarce and shallow attention.

And the fact that they’re teaching it to lie under cover of ‘alignment’ only means that the research into telling lies is progressing faster than research into telling the truth with some degree of logical consistency.

Cheers.
Sorry to counter-signal but I can’t find a use for openai or any other group’s work so far, other than code completion.