Since 85% of Mozilla's current revenue comes from Google paying to have their search engine set as the default on Mozilla, and given that this is at risk due to last year's antitrust ruling against Google, they will need to diversify regardless.

Focusing on removing unproductive activities sounds like the best approach, as it feels much easier to identify them than to weigh the real productivity ofany task. But, as Goodhart’s Law says, "When a measure becomes a target, it ceases to be a good measure."

Not really in touch with this type of article. Comparing tools seems a bit meaningless to me; I prefer to focus on the engineering side of being an AI engineer — what they built, how they built it, and what AI brings to the table compared to traditional programming.

Very cool work! And I love the nuance in your methodology and findings. Anyway, I'm preparing myself for all the "Bombshell news: AI is slowing down developers" posts that are coming.

Well, it would be a nice counterweight to all the gaslighting of people who claim AI doesn't improve their productivity...

I've never seen a post telling someone who claimed that AI didn't improve their productivity that they were mistaken and that actually it had.

Plenty of "you're holding it wrong" and "it works for me", but not directly telling someone that their perceptions were inaccurate.