Charlie Loyd @vruba@everything.happens.horse

Had a lot of fun with @salkjorts on the other site, where I posted bad ideas run through Midjourney-or-whatever, so please welcome @salkinjorts. Or block it, if you don’t like this stuff. I’m seeding it with a few redos of classics from the old account.

I continue to feel pretty alone in the interpretation that these models are neither trivial nor conscious; that they are neither nukes nor penicillin; that practically all the urgent questions about them are really the same urgent questions we’ve had about political economy, offshoring, and corporate governance for decades; and so on. I finally get to be a filthy centrist on an important issue and it turns out it’s boring.

This statement combines a lot of what I think is good and less good about the thread of LLM criticism it’s from. I don’t think the stochastic parrot interpretation is more than one interesting model, I believe the relevant problems with longermism are almost completely disjoint from the ones mentioned, and I am unconvinced that marking media by consciousness of author is either philosophically clear or practically useful. However, I agree firmly with the tl;dr. https://www.dair-institute.org/blog/letter-statement-March2023

If you know where to get originals of these terrestrial-but-field MSS images, I strongly encourage you to share.

I mention this like once a week, but since she’s just died, let’s all go look at the ~20 cm Landsat image of Half Dome that Virginia Norwood made.

«Norwood had researchers load a breadboard version of the scanner onto the back of a truck. It amounted to “just a bunch of boxes,” she says. “We could use all the weight we wanted.” […] Having labored so long over the specs, Norwood wasn’t surprised by the high quality of the test images.»

https://www.technologyreview.com/2021/06/29/1025732/the-woman-who-brought-us-the-world/

Sometimes I think about learning Unity or Bevy or something and dedicating several years to doing this the way I want it done. Then I remember this is a spectacularly terrible idea. But I do have a text file of funny species names in case someone does.

I would spend my entire yearly entertainment budget on a modern, richly implemented, beautiful, highly replayable version of SimEarth (1990) or SimLife (1992).

Someone decided in 2018 or so that all “friendly” software would be eggplant-colored and I just want to ask them a few harmless questions.

Coming up on 70 years of person A going “You see, software will never be able to do X!” and immediately being proven grotesquely wrong, then person B going “You see, this software is basically a person!” and immediately being proven grotesquely wrong, and so on like the world’s most evenly matched and annoying game of table tennis.

What’s the more familiar, less lit-class way to talk about Keats’s negative capability? Intellectual humility? Tolerance for uncertainty? The ability to begin without knowing or even having faith in the end? I need more of that and so does the world.

It would be extremely funny if the big social media companies’ lobbying and PR* against Tiktok is what gets us a real privacy law.

https://mastodon.social/@jameeljaffer/110084814863743927

* Think about how many stories you’ve seen about Upsetting Tiktok Trends from outlets that don’t normally cover social media. That’s PR spoor. And the current dominant platforms’ predecessors did it to them.

Another exhausting and lonely day in the being right factory.

I’ve closed a lot of tabs of “well, obviously cognition is computation” or “well, obviously cognition is not computation” or “clearly, truth is/isn’t XYZ” or whatever. These are far-reaching arguments that depend in super nuanced ways on a whole shelf of context. They’re weighing you down, not helping, when you want to say is “merely” something like “this model is/isn’t useful in that context and will have these larger consequences”.

Free scare quotes to distribute as you see fit: “ ” “ ” “ ”

Perhaps there is a narrower way to make your case about what the software is doing, and it might make your case more convincing to people who don’t happen to subscribe to that particular grand theory of how cognition works.

Maybe a better way to say the thing I’m trying to say is that a whole lot of people who feel a desperate urge to tell the world (1) “ANNs are better than us and also the new bitcoin, which is good” or (2) “actually, ANNs actually aren’t even real, actually” is: If you start by tying your sense of what a piece of software is doing to a massive, categorical, totally abstract philosophical idea about the nature of being itself, maybe go outside and look at the stars for a minute.