GitHub finds 92% of developers love AI tools
And said tools are therefore less popular than Saddam Hussein, who was elected with 99% of the votes.
I both cases I have questions about the methodology.
GitHub, which has been doing a brisk business selling subscriptions to its litigation-encumbered Copilot AI helper, surveyed 500 US-based developers to find if any of them are already using AI coding tools at work. As it turns out, they are, and in a big way. We know, you're shocked. The Microsoft-owned code storage biz …
“ Asked how they are currently judged, these programmers responded: code quality (40 percent)…. etc….”
Well that sums up nicely what’s wrong with modern software development. What a load of slipshod useless f*$€rs - I would sack the other 60% on the spot.
Of course, it’s more fashionable to “release often and break things” - what a f*$€ing stupid mantra - and I’m an old git that uses vim and writes code that works (not that it counts for anything)
That doesn't have to be their fault. I'm not sure I've always been judged on code quality. If so, there are some places I've worked where some people were not being judged correctly. Sometimes, management really does prefer the completion of whatever tasks they just set, regardless of whether those tasks are completed well or if they were worth completing.
I find that the latter tends to come first; someone is given a task that has no purpose and management wants it done fast, so any time they spend on doing it well feels like time wasted and is punished by managerial complaints. It's not surprising that motivation is lower for task number 2, leading to lower quality. The survey also indicated that progress was measured in lines of code, which every manager should understand is a bad measure of capability or productivity, so we're not looking at well-run projects here.
There is probably also a sample problem as well; my guess is that GitHub managed to get 92% agreement on AI tools by putting the survey invite in something that only people who had used Copilot would see. They filter out individuals and small companies later, and 8% of the respondents were those who tried out the tool out of curiosity and didn't like it. I can't prove that, of course, but that's how I would expect a 92% positive survey result to be generated if managers indicated that they preferred that result to honesty and survey quality.
@Rich 2 "Well that sums up nicely what’s wrong with modern software development. What a load of slipshod useless f*$€rs - I would sack the other 60% on the spot."
Why? The programmers are not responsible for the metric their work is judged on. If you sacked the 60% and employed new programmers there would still be 60% who are not judge on their code quality. Just because a programmer is not judged on code quality does not mean they do not write quality code.
The 60% of programmers are not the problem, the problem is the employers/management that have chosen not to judge on code quality.
> The 60% of programmers are not the problem, the problem is the employers/management that have chosen not to judge on code quality.
OK, so at what point as an "industry" of developers do we take responsibility for our work?
How come when anything about the fucking horrific quality of code shat out by our no-professional-standards "industry", both commercial and open source - we somehow always, always get to blame someone else?
Take responsibility for the quality of your work.
"OK, so at what point as an "industry" of developers do we take responsibility for our work?"
At many points, but not the one you're complaining about. If your management judges you on something, you don't get to choose that. It might not even affect your work. I've written good quality code despite management clearly not caring about the quality, just wanting something that works. So if the statistic shows that 100% of coders are not judged on code quality, it wouldn't automatically follow that 100% of code is of poor quality as some will be self-motivated to produce good code. But even if that's not the case, this is clearly a management issue. Let's take an example from an unrelated field:
Let's say that you work in a call center and I'm your boss. You're receiving calls from customers who have problems and you're trying to solve them. I've decided that these customers are highly likely to remain customers, so there's not much need to treat them well. Therefore, I've given you instructions to go through a basic script to fix the really easy problems, and if it's more complex, leave them on hold forever until they give up. Since I've given you such a simple script, you need to spend no more than three minutes on a call, and if you do, I'll be monitoring and harassing you over it. Whose fault is it that the customer service is terrible? If we're following your logic, it's your fault. If we're following common sense, it's clear that it is my doing that led to the situation and it is I who need to change if the situation is to be resolved. Blaming people for management's mistakes not only punishes the wrong person but prevents you from improving the situation.
I've seen people come in with "AI" generated/assisted code. It looks like it might work on the surface, but then a lot of time is spent figuring out why it doesn't work or how to make it secure and appropriate in our environment. At that point, you've spent more time trying to force the 'lazy' way to work than you would have if you'd just done it yourself from the beginning.