zerd

Adjust brightness/sound: I can do that very easy and precisely without a touch bar. There's a button in the top right corner that has a slider. Both take about the same amount of time/movement to do.

My main gripe is that some of the buttons move around depending on which app you have focused.

They're going to force us soon. When getting a new device I don't think you can install an older version either.

Interesting, I made a Claude code plugin to do just that two weeks ago.

Claude just used the word “ Netowrk” instead of Network. Mind you, I was only using Sonnet with reasoning, but I thought trillion-parameter models would be able to spell a common word.

Gemini is the worst gaslighter. It hallucinated a feature in an open source product. I said no that does not exist. It invented very realistic git commit messages. I said commit does not exist, so it said I was out of sync. I linked the repo and it said it might be a fork. Then it invented a mailing list thread where they supposedly removed the feature, that’s why I couldn’t find it.

Searching for anything product related is now AI answer, followed by “Sponsored products”, followed by “Popular products”, a few hits, then “Shop by store”, and “In store nearby”. Maybe 2% is actual content, the rest is ads.

Their prediction was spot on: “We expect that advertising-funded search engines will be inherently biased toward the advertisers and away from the needs of the consumers.”

Limiting it to just read and http limits its usefulness. If you want something like “filter for this, count the number of matches, format like this” you have to make custom tools. And you have to make sure they don’t have any bugs that allow arbitrary code execution. You’re effectively building your own sandbox in this case. Using a standard one sounds easier unless you have a very focused use case.

Time to start some new restaurants matching those. Like Bubba Gump.

How do you scale past 16 channels though? 16 channels give you around 614 GB/s, while PIM can do that per chip, so it can achieve 58TB/s.

I’ve been doing the same thing, giving the same tasks to Qwen 3.8 27B and Opus, and the main difference is that Qwen does not consider edge cases which Opus catches. It’s good at the happy path, but even when hinting that there are uncovered edge cases and gotchas it’s oblivious to it. So I feel like I need a bigger model to do planning/review.

"We're doing it for your own good, pinky promise"

I noticed Claude code does something similar when "Exploring" the code base, it spawns a subagent using Heroku to get the summary and file locations, which is both faster and it doesn't pollute the context as much.

I wonder how it compares to an vector indexing approach.