Benj Edwards at Ars Technica analyzes Anthropic’s announcement of the latest version of its Claude AI, which for the first time beats GPT-4 on benchmarks and demonstrates “near-human” capabilities in some areas (or so Anthropic says).

Benchmarks don’t necessarily show how effective the tool is, Edwards notes.

Also:

It’s probably true that Opus is “near-human” on some specific benchmarks, but that doesn’t mean that Opus is a general intelligence like a human (consider that pocket calculators are superhuman at math).

I used Google to find a PDF of the movie “Love, Actually.” I fed the URL to ChatGPT and told it to create a poster. It said it can’t access documents on the Internet—even though it had done so a few minutes earlier. I uploaded the PDF and tried again. It gave me a text description of the poster. I told it to create the poster and it did.

Lazy robot!

I used Google to find a PDF of the screenplay of “Die Hard.” I fed the URL to ChatGPT, with the prompt, “Here’s the screenplay for a movie titled ‘Die Hard.’ Create a poster for this movie.”

I asked ChatGPT to create a poster to promote the TV show “M*A*S*H.”

City leaders in Escondido reject “Housing First” policies for homelessness because they say homeless people need addiction and mental health treatments first.

“Homelessness is a complex problem that requires complex solutions,” San Diego County Supervisor Jim Desmond said.

I’m skeptical. Are most homeless people mentally ill? Or is the problem caused by housing being too expensive and scarce?

And even if many mentally ill people are addicted and/or mentally ill, to what extent is homelessness driving that problem? Living on the streets could drive anybody nuts, and make them turn to alcohol and drugs to get through their lives. And it seems ridiculous to ask someone to get sober and sane while they’re living on the streets.