[uptime]

I did not enjoy this - you need to get all the way to the end to see that he is ok with silos as long as his get to survive. The article is a bunch of breadcrumbs to get you to the point where he springs his models on the reader.

Which is would have been one thing he if was honest but the frame is bogus. If he wanted real diversity at the popular level he would not have cherry picked so very hard so assign blame.

His frame is that ‘wokeness’ is the real problem with AI, since results from AI language interfaces will be governed by whatever that is. But all sorts of systems can present the problems he wants us to understand. You can’t tell me that Thiel/Zuck/Q/4chan or any other entity on any spectrum could not pose a problem of similar shape, depending on ownership or algorithm mastery.

Having a bitcoin guy tell us the free market is gone and info homogenization will come because commie cat ladies, with no mention of megamergers and market capture through non competitive means, seems like he is burying his own values just to print some dunks he thought was clever.

It is rare for me to more insulted by other terms in a text when the author uses the word ‘ideate’ non-ironically but it happened here.

I am eager to see good debates about language mediated interfaces and debates whether LLMs will expand or contract knowledge, but not starting from his set of assumptions.

[drvdevd]

I think I agree with you on how I felt at the end of the post, but at the beginning and most of the way through, I found myself generally agreeing. But I want to look past the “bitcoin” switch after the bait because I generally agree with the author on most points.

Especially as I have delved more into how LLMs actually function and are trained, a lot of the magic in e.g. ChatGPT has evaporated for me, but in a good way. The performance of larger models like GPT-3.5 or 4 is still consistently impressive, to me, but as I’ve increased my understanding I do believe that censorship poses a greater threat than the models themselves.

[uptime]

Yes - and I would expand “censorship” to “control”. We have been shaped by whomever has built the algos for google, fb, twitter, etc and while that itself doesn’t make them evil they need to be evaluated somehow. I think the LLM fear comes when the creators tell us a full evaluation is impossible, etc.