Man, I was ahead of the time on cancelling my ChatGPT subscription and going full time Claude!

So the government turns its back on Anthropic for disallowing two use cases and then turns around and makes a deal with OpenAI that disallows those same two use cases? Are the people in charge of our government total fucking morons or what? Yes, that’s a rhetorical question.

I enjoyed listening to Jason Snell get "rascally" about how the Tahoe hate bandwagon is extremely overblown. I don't think Mr. Gruber realizes that he's been super negative about Apple ever since they shunned his WWDC talk show episode last year, and that that may be motivating his constant bloviating about how awful Tahoe is, despite not using it. Gruber's shit talking on Tim Apple is justified, but I think it is also maybe a little turned up since Gruber stopped being Apple's favorite little fanboy.

Recently a friend of mine, who is undoubtedly an expert on the topic of which he was speaking, said, "Your current bot is better than most of the other free stuff I use". So... there's that. I've got that at least.

i heard a new one today from the realm of "agentic superstition"... a friend told me his coworker swears opus 4.5 was a downgrade for claude code, now they use claude code to plan an implementation, then they take the implementation doc over to Codex to have Codex write the code.

i have to wonder what is going on in the heads of people like this, the people who fill reddit every day going "x LLM is bad today! x LLM gives me better results at 1am than at 12pm! x LLM used to be so much better just days ago!", what are they doing?

Due to the approaching Discord age-verification apocalypse, for fun I set up a Discourse instance at discuss.bradroot.me to possibly serve as a replacement for my Discord server, which no one really uses anyway.

A winner emerges... I've cancelled ChatGPT Pro and Perplexity Pro... I am only paying for Claude Max now.