Brain worm, question mark?

Don't forget a quick plug into the "FO" port.

Brought to you by folks whose holy grail is the live service model.

"Legalized" extortion.

Agree on the rubber ducky. Also saves a bunch of time when solving an odd edge case. For me, the code completion is just meh, as it tends to make enough mistakes that I end up reviewing whatever it creates, often nullifying any time savings.

Copilot/ChatGPT/etc. is useful when I need to look up certain implementation details, or even get some higher-level guidance for certain topics, but I've been in the industry long enough to know it's not the panacea the hype machine likes to pretend it is.

"Useful" is not, and has never been, their end goal.

I've been waiting for this since the pictures.

I can appreciate Azure Cosmos' model of RUs/sec, which also quantifies the amount of work rather than strictly call volume. Implicitly, it's directly correlating work to cost, and on a time/capacity basis rather than overall usage. That said, as the caller it's not always easy to forecast.