My feelings about AI are weird. On the one hand I really do not like the amount of slop it generates (such as in creative spaces, videos) and enables and on the other I can see it's geniune usefulness as a tool. "What's this error mean?" "Parse this log for me and get the data I need", etc. I see the usefulness in that.

Part of the reason I don't like it is the amount of bullshit it generates in terms on traffic on the internet. I see it every day at my job, normal peoples sites being scraped and hit just so the AI crawler can do whatever it wants and ignore robots.txt. It makes people that normally wouldn't have to deal with things like cloudflare and makes them use it just to get their blog working. To me that's very stupid.

I think what AI needs is oversight. We need to stop over building datacenters just to give it more power and we should be working on making it more efficient. I don't really know how much we have left to give in terms of computing power due to moore's law, so unless recent research becomes mainstream (diamonds transitiors) I don't know that it's going to get much better so I think we probably need to just accept it's as good as it's going to get. I am probably very wrong on this.

I have seen people that normally would think for themselves and just verbatim accept what AI says as fact. Can we not do this? AI needs to be reviewed constantly on it's output.

Linus Torvalds recently spoke about it's use in the linux kernel and I do agree with him somewhat. There is no way that any maintainer would let slop into the kernel, that's just not going to happen. Fork it if you don't want AI in it (I do kind of want someone to do that) but it's existance in the kernel isn't the end of the world, it's objectively going to be fine, linux isn't just going to get awful now.

I dont know, I'd say overall I don't particually like ai and kind of wish it never existed, but we're here now and have opened pandoras box so it's probably pointless to not at least use it as a tool.