I vastly prefer digital methods of food ordering. I don't really care if it's delivered, but the last thing I want to do is take up some overworked food service employee's time while I rattle off any modifications I want/need to make to cater to my food sensitivities.

the hong kong protestors used the phrase "do not split" to mean that the radical protest and peaceful protest arms of the movement should refuse to criticize each other.

Considering we're going back to Racoon City, this is more likely than ever!

I don't really find the "Electric District" line to be that much different from the others and it's - obviously - electrified. Maybe it has more to do with smoothness than slowness? I find the ED much more pleasant of a ride than the L, for example.

If we're asking Dems to pick between those things, we've already lost.

They should be doing all of those and, if they don't have the wherewithal for that, they shouldn't be in Congress.

I recognize that I have strong, negative opinions on generative AI (and LLMs in particular) that are coloring my perspective here.

Overall, though, I find most common uses of LLMs to be net-negative. They do have positive uses, though, and I don't want to completely ignore that point.

I would argue that kind of anti-social behavior is a net-negative to humanity and we probably shouldn't be promoting or rewarding it.

If ChatGPT was a real person, I would want absolutely nothing to do with them.

Fun fact: I worked with Joel Berez at my last job. One of the best engineers I've ever worked with.

This does make me wonder if it would be possible to coerce an LLM into being an effective engine for a game of those types. Might be interesting.

I disagree with your assertion that it approaches understanding, but I feel that's a matter of philosophy. If an LLM was produced that could not trivially be made to hallucinate basic facts, I would be much less confident in my disagreement.

I don't understand what's so bad about pay-by-plate or transponder-based solutions? Is using an app or a website really that much worse of an experience?

To my knowledge, LLMs don't operate like this. They take some prompt and, using that context, effectively ask "what word comes next?" until it reaches an end point.

LLMs are fairly good at giving the appearance of understanding, but that's all it actually is.

A good answer to a bad question @katmabu.bsky.social www.readtpa.com/p/absolutely...

This sounds like the same "we don't need EVs; we need public transit" argument - letting perfect be the enemy of good. I don't care where the houses get built _just build the houses_.

Do we need more dense housing? Absolutely, but not everyone wants that and we need to accommodate them as well.

And, with one exception, all the places I did stop to charge were in very nice, convenient locations where I could go eat, take a walk, or use the restroom. All things I would have had to stop to do anyways.

I took a road trip to Kansas City then on to Tulsa in a Tesla once a few years ago. If the rental hadn't been missing it's plug-in charging adapter, I would have needed to stop _twice_ to fast charge the car in total.

That's _so much less_ time than I would have had to stop with an ICE car.

If I remember correctly, this also supplies a cooling effect, yes? I don't recall it being very much, but even a couple of degrees of cooling can help when it's very hot out.

Your question boils down to "what if the president had run on the D ticket?" The correct answer is always "fuck fascists".