

I don’t think it’s fair to compare LLM code generation to machine vision in this way. These are very different "AI"s. Not necessarily disagreeing with Doctorow, but this is an important distinction.


I don’t think it’s fair to compare LLM code generation to machine vision in this way. These are very different "AI"s. Not necessarily disagreeing with Doctorow, but this is an important distinction.
Kolmantena päivänä 10 000!
I came here to say this. Glad to see that it already has been said. The economy is bonkers.


I was hoping the author would advise against using this, because after doing this your website isn’t accessible.
I don’t think this is the answer to web scrapers, since it so adversely affects people with disabilities. The web is one of the few things that is structurally (reasonably) accessible by default even if making inaccessible websites (like the author’s is now) is possible.
If you have to choose between accessibility or letting scrapers to access your site and scrapping accessibility, you should choose accessibility.
I upvote when I think other people should see the post as well. I downvote when the content is bigoted or similar and doesn’t add to the conversation.


I don’t think you mean “1 lb of protein”. I think it’s a around 0.5 grams per 1 lb?


The article doesn’t say they do this. It says: “In theory, a service could charge you more if you’re on a vacation in a foreign country, instead of your hometown, because they know you would be willing to pay more. It would be the same if your phone’s battery is low and you’re trying to get home.”
A really bad headline, but not surprising I guess.


It’s on a good roll now. People should go and sign!


“Oh my god! It’s a giant rock!”
“Oh my god! Another giant rock!”


Bssed on WebAIM Million report, it could be said that it has improved marginally (https://webaim.org/projects/million/). Yes, maybe the tools that people use are getting better, but the native experience is not getting better a lot.
From the report the biggest accessibility pitfall is still poor contrast. It’s sad that user would need tools to access information that only has poor contrast.

Would you mind sharing what kinds of things are taking up your time?
Seconded (from Finland). Math teaching is still boring and it doesn’t help that (too) many teachers don’t appreciate pedagogical studies that go with the curriculum in the university. It has its problems, but rarely do the students want to engage with what math education could be.
Coupled with society’s expectations on what math education is, it’s really difficult for a teacher to change course. Even the students have the expectation that they should always be doing exercises from books and everything else is ‘useless’. It is really a deep rooted issue.
One addition to this is also winter upkeep, which is very relevant in Finland.
People like to talk about “winter cycling”, because it’s somehow so much different from “every other season cycling”. Mainly it comes down to winter upkeep; snow plowing and such. Then some people complain how nobody rides in the winter and they shouldn’t use too much budget for it.
It would be fun to see people talk about “winter driving”. How much we actually spend making driving possible during the winter.
YO-kokeilla on sama vaikutus. Yleissivistys saa väistyä sen tieltä, että pääset kokeessa “keksimään itse kysymyksen ja vastaamaan siihen”. YO-kokeet roskiin.
Sounds like they had it coming. What are HTML boxes and HTML fields?


This would be interesting to me as well!
My understanding is that the tumor detecting machine vision is generally thought useful in addition to the radiologist’s expertise. It basically outputs “yes”, “maybe”, and “no”, which is more expertise respecting than generating somewhere thereabouts code, which the coder has to (now) validate.
This is why I wouldn’t equate these tools. LLM code generation is marketed to do much more than machine vision for tumor detection.