I have decided not to post it. The discourse around Dack has gotten too heated, and I don't want to fan the flames. I made top 16, ended in 7th/66. I had a really good time, some of the most fun magic I've had in a while. But I'm just going to leave it at that. Be good to each other, it's a game.

Show Post

And this is what I've been saying the whole time. The initial gains are lost because a good dev will review it anyway and then lose that time. But you also then have a dev who then hasn't developed new knowledge from figuring out how to build the software themselves. It's a net loss overall.

Show Post

AI-driven code reviews are unreliable as any point it gives you has to be investigated, and in the end is no substitute for a dedicated lead on a project. It's also no substitute for a lead to be more hands off because then in a review time is wasted giving that lead the context AGAIN

Show Post

I made a big arse out of myself last week because I was adamant something worked a certain way, and it turned out after a while I was introducing the bad test data. Lead said "it's not good enough" Copilot makes way more mistakes way more often. Why does that get the free pass every time?

Show Post

And again, it's all under the assumption that these models are smart and technical. They are not. They are text generation tools. That is still all they are, they are not trained specifically on how to build or engineer software. You have been fed marketing lies and you've believed it all.

Show Post