[bigDinosaur]

Worth noting that copilot has been in active use by many devs I know before ChatGPT and that doesn't appear to have eliminated jobs. Fundamentally your job is to understand - not just type characters into a terminal. I think you might be shocked how bad things would get for a person who lacks any understanding of what they're doing if they just concatenate ChatGPT output indefinitely.

[SOLAR_FIELDS]

I'm not sure how other people use Copilot, but the only way I use it is "Really good autocomplete". Like if I'm typing out 3 repetitive lines, the fact that it figures out the next one for me is amazing. I don't have it writing entire unit test suites for me or designing approaches like with ChatGPT - so I wouldn't even call them in the same league in terms of comparison.

[ipaddr]

It worked great for suggesting seed data

[simonmesmith]

Agreed, but I’m not just talking about engineers. From conversations I’ve had, I think it’s more a question of what will be possible with generative AI over the next year, and whether that uncertainty makes it prudent to hold headcount steady or let it decline. The uncertainty includes the possible impact of the tools on automation, and also how much the tools might solve problems that entire companies now exist to solve, negatively affecting sales. For Copilot specifically, if it enables developers to be even 25% more productive, and you’re planning headcount, do you reduce your targets by 25%, or raise your ambitions?