There isn't a lot that gives this old man hope these days, but this does. More of this, please.
There's more data than ever to work with, and someone will do the work. The key is whether the work is done for good or for other reasons. When we project the impact of AI on this, it starts to look like a first past the post winner take all situation, at least for long enough that the impact will be exponential for the good or for the bad.
Addendum: It has come to my attention that the author of this proposal, Carlos Hoyt, has a working relationship with two individuals I highly respect. Those would be Greg Thomas and Sheena Mason. The proposal's update with editorial on 'reparations' has been removed, and thus my prior objection. I urge you to sign the petition at Change.org (https://bit.ly/3SOz5FS).
Yes. AIs do it without explaining their work. But I think the aegis of what's going on in AIs is something that is well-known. I've got a lot of learning to do, but I expect I'll have better answers in a couple years. Right now I'm just starting to take data science more seriously, which may sound ironic given my career, but I've never really trusted the newbies and the Google bros. That's primarily because they haven't been interested in serving the purposes of reforming academia for one thing and solving various important oracle problems for another. Plus the amount of FUD and scamming in crypto, not to mention wishful thinking, is massive. But that doesn't change the fact that I've been following Balaji Srinivasan and Vinay Gupta and I think they have massive ideas. More on them later.
Thank you for that lucid discussion, and for those names -- I'll follow up on my own. I'd like to cut through the balloon juice around these topics. On a lighter note, your mention of imputation took me back to my dissertation (shudder...PROC MI in SAS 9).
There isn't a lot that gives this old man hope these days, but this does. More of this, please.
There's more data than ever to work with, and someone will do the work. The key is whether the work is done for good or for other reasons. When we project the impact of AI on this, it starts to look like a first past the post winner take all situation, at least for long enough that the impact will be exponential for the good or for the bad.
Godspeed. I hate interesting times.
Addendum: It has come to my attention that the author of this proposal, Carlos Hoyt, has a working relationship with two individuals I highly respect. Those would be Greg Thomas and Sheena Mason. The proposal's update with editorial on 'reparations' has been removed, and thus my prior objection. I urge you to sign the petition at Change.org (https://bit.ly/3SOz5FS).
I find your substack interesting and “off the beaten path”. I’m not a data person or a numbers gal, but I appreciated reading this. Thanks!
Yes. AIs do it without explaining their work. But I think the aegis of what's going on in AIs is something that is well-known. I've got a lot of learning to do, but I expect I'll have better answers in a couple years. Right now I'm just starting to take data science more seriously, which may sound ironic given my career, but I've never really trusted the newbies and the Google bros. That's primarily because they haven't been interested in serving the purposes of reforming academia for one thing and solving various important oracle problems for another. Plus the amount of FUD and scamming in crypto, not to mention wishful thinking, is massive. But that doesn't change the fact that I've been following Balaji Srinivasan and Vinay Gupta and I think they have massive ideas. More on them later.
Thank you for that lucid discussion, and for those names -- I'll follow up on my own. I'd like to cut through the balloon juice around these topics. On a lighter note, your mention of imputation took me back to my dissertation (shudder...PROC MI in SAS 9).