AI, lol

Well, I said I'd ramble about AI. Certainly I don't know anything about it. Like many folks, I'm optimistic and also a bit petrified. From a young age, I wanted to be a computer programmer. That, and I wanted to collect information as a way to be useful/employable. Need to know how to do something on the computer? I bet you I've got a plan.

Nowadays with AI and the advent of even better search engines, such a self-concept is errant. Being able to scrape together a good script (the extent of my programming abilities, arguably) is pretty moot if you can just ask the AI to build it, and faster, and arguably better. And in the same vein, you may as well just ask it what you were going to ask me. If you don't, it's likely I will.

When generative AI was starting out, I honestly didn't believe it. Images would be posted to social media saying it was made by AI and I figured "nah, computers just can't do that yet. Don't bother." It was another two or three years of this before I finally opened up the AI tool and asked it to build an image. Granted, the quality sucked, and/also I didn't/don't pay for the services. But, I had to face 1) I had become the sort of curmudgeon that doesn't keep up with technology or accept its progression, and 2) it was in fact progressing.

It took me a couple years after that before I got particularly comfortable with using AI on a daily basis. For me personally, I don't see it as terribly helpful in the stressful moments of my work. When I /need/ help because systems are misbehaving, or acting unpredictably, the AI is of little help, or honestly a bit destructive. It'll ask for information that does no good, try to solve irrelevant problems, etc. So I know that to this point, I cannot truly "rely" on AI.

But, I do think it has its uses for me. I'll pull it up for a sanity check on something. When I was looking at changing spark plugs or other car maintenance I'd never done, I'd lump it into my research options. Find several YouTube videos, ask a generative AI what ideas it had, then take a synthesis of the two. Works for me.

Of course, it still gets stuff pretty wrong, even for mundane/not stressful things. (Sidebar: I had a plant shipped from a nursery. Came just fine. But it was actually two stuck together. Well, one got the better hand, the other started dying. What do I do with the other one? Snip it at the ground? separate the roots? How big of a risk is rot from a slightly dying plant???)

Anyway, I used to think that AI was completely useless and risky for bad information. And it is... but also I've grown accustomed to being lazy. When Google makes a conclusion at the top of the search results, I'm finding myself actually just reading it anyway and taking it as is. Has the service just improved, or has my criteria just dropped to accept it? Dunno. I don't think I'll ever pay money for the services, though. I'd sooner go with some open-source option and pay a higher electric bill than motivate that economic shift.

But at the end of the day, I think what draws me to the generative AI is truly that "personal" touch. (No, not "personalized", as I don't even log in. My conversations don't need to be stitched together as they're purely for my edification.) But the computer can now say "sure, that's a good idea" in its obsequious manner. Or, it can facilitate my curiosity on obscure concepts that we bloggers may not have really focused in on specifically, but that can be groked together through the synthesis across the internet. Perhaps even less tangibly, I like being able to take my rant, my confident stances, and just throw it at the computer. In terms of things like macroeconomics, the computer is pretty quick to tell me how wrong I am. And that's just terribly cool. But even more than just evaluating my claims, my questions, and so on... it's a mechanical ear. I don't have anyone I can just talk about these concepts with at length, learning or comparing notes. And I'm not "mature" enough to recognize that maybe such interactions only have value when they're with other humans. For now, I'm content screaming at the elctronic non-void.

I had glanced over another blogger's comment on Plato's Republic. There was a section we really sped through that seemed (to me) to indicate that maybe your better philosophers can get their start by having lived more of a hermit lifestyle. Away from the general beliefs held by the public as a whole. I asked ChatGPT, and it says I'm pretty off-base on that.

One last note, I still think that Google is sandbagging their own search results to make their AI look better + improve adoption. I have no proof, I'm pretty sure it's not libel to claim it. But man their search results suck lately, but somehow their AI "assistance" can find things better. Bearing in mind that the line between "generative AI + complex algorithm finds your webpage" versus "complex algorithm finds your webpage" feels like a relatively inconsequential leap to make. In other words... they had the technology, and they seem to have it now, but somehow it doesn't translate into good search results. Dunno.