I think it was one of those situations where everyone just pretends they don't notice what's happening.

This wasn't the case where I grew up. Woodworking teacher regularly drunk after lunch. Alcoholic principal. Though I suppose the 90s were a different time 😅.

This sounds like a winning startup idea to me.

This is what makes so much LLM "research" now such a joke. The big labs have no idea what training data is going into their models. So they run these tests, but can't tell you if the answers to the tests are in the training set.

Guess I'll see what shows up and rejoice either way ;)

Seriously? I just ordered one on Friday 😅

If you make it out to Vancouver we've got a spare room you're welcome to crash in :D

I was driving home (5h drive) and asked Gemini to navigate to a specific gas station and it suggested that i should pull over and rest if I'm feeling sleepy. It did do the navigation thing at least.

There are plenty of things to be critical of ai gen code, but these arguments are always a moving goal post. In January it was "where are all the new GitHub repos" in March it was "all the GitHub repos have no stars". Now, no one uses the apps. Etc.

Thanks! Just went back and registered the domain I've had in my cart for 2 weeks.

already sold out!

@trueberryless.org The repo is tangled.org/checkmate.bl..., the basic app is workink pretty well, but I'm also working on some server-side stuff with the atchess-bot to keep track of games and stats. Hoping to find some time in the next month to give it some more love :).

Dropped a skill for it here pdsls.dev/at://did:plc... it will want the atproto-records skill as well.