m_w_

I don't understand this comment at all - it's very obviously written about personal experience with development, doesn't read like it was written with AI, and the sentiment is nice.

Sure, the projects mentioned aren't the most impressive pieces of software ever written, but isn't that kind of the point of the article?

> Presumably, these drugs are (ridiculously tightly) controlled to prevent society-wide harm.

Maybe I'm misunderstanding what you mean - but I think almost any college student would disagree with this presumption.

> Do you have a source for the benefits giving way to problems long-term?

Although a very long read, I found this to be very insightful:

> It was still true that after 14 months of treatment, the children taking Ritalin behaved better than those in the other groups. But by 36 months, that advantage had faded completely, and children in every group, including the comparison group, displayed exactly the same level of symptoms.

https://archive.is/20250413091646/https://www.nytimes.com/20...

I can't speak to the content of the actual game being played, but it wouldn't surprise me if there was an in-game text prompt:

> "The house that looks like a ripe tomato!"

that was transformed into a "user prompt" in a more instructional format

> "Go to the tomato house"

And both were used in the agent output. At least the Y-axes on the graphs look more reasonable than some other recent benchmarks.

Glad the snippet is == 1, this problem drives me extra insane when it doesn't handle the zero case. Not only has my upload failed for some mysterious reason, I have successfully uploaded zero item!

Homepage layout horizontally overflows <1000-ishpx wide to when it switches to mobile layout. Cool concept though.