This is such a weird poll. How much you use AI is a separate question from how much more productive it makes you. Trying to put the two together in a single question, and then only allowing a a few answers, means you give no flexibility to actually answer the question outside of your expected results.
The reasoning is that you would have productivity gains if the AI actually generates the code you need. If it doesn't, you either spend your time writing most of the code, or else spend significant time debugging it. Maybe you're trying to coerce the AI to correct the bugs for you and it doesn't work out. So your work may theoretically have changed to 100% AI, those debugging cycles only negate the productivity gain, so the 100% figure is not real.
> No, the AI does it all and I'm 10x as effective. Hardly any polish is needed.
> No, the AI writes it all. I'm 3x as effective. It still needs substantial polish.
You're only allowing people to say, "AI does it all and I'm more effective because of it." You don't even admit the possibilities that AI is doing it all but people are only as effective as before or that they're less effective. Your "yes but less than 100%" answers similarly don't offer sufficient variations. Why is 75% the only break-even option (when using AI) and not 50% or 25%? Below 75% you only let people say "AI is writing at least as much code as me, and it's made me more effective".
This poll needs at least two axes: What percentage of your code is written by you (versus AI)? and What effect has it had on your productivity/effectiveness as a developer?
HN polls are ill-suited to this unless you want to make 25+ polling options.
This is such a weird poll. How much you use AI is a separate question from how much more productive it makes you. Trying to put the two together in a single question, and then only allowing a a few answers, means you give no flexibility to actually answer the question outside of your expected results.
The reasoning is that you would have productivity gains if the AI actually generates the code you need. If it doesn't, you either spend your time writing most of the code, or else spend significant time debugging it. Maybe you're trying to coerce the AI to correct the bugs for you and it doesn't work out. So your work may theoretically have changed to 100% AI, those debugging cycles only negate the productivity gain, so the 100% figure is not real.
Look at these two statements in this poll:
> No, the AI does it all and I'm 10x as effective. Hardly any polish is needed.
> No, the AI writes it all. I'm 3x as effective. It still needs substantial polish.
You're only allowing people to say, "AI does it all and I'm more effective because of it." You don't even admit the possibilities that AI is doing it all but people are only as effective as before or that they're less effective. Your "yes but less than 100%" answers similarly don't offer sufficient variations. Why is 75% the only break-even option (when using AI) and not 50% or 25%? Below 75% you only let people say "AI is writing at least as much code as me, and it's made me more effective".
This poll needs at least two axes: What percentage of your code is written by you (versus AI)? and What effect has it had on your productivity/effectiveness as a developer?
HN polls are ill-suited to this unless you want to make 25+ polling options.
The previous polls:
- 30 days ago: https://news.ycombinator.com/item?id=43584842
- 60 days ago: https://news.ycombinator.com/item?id=43265234
Don't forget to upvote, because previous polls didn't make it out of the `new` section.
> Recent advances in AI promise to make most of the handy work of coding a thing of the past.
Talk about begging the question!