It seems screamingly obvious after nearly a year of intense hype that the "I" in "generative A.I." is not in the tools themselves, but in the data they were trained on. And that's their biggest weakness: ask 1,000 developers, get 900 bad answers.
Nearly 6 months after the first person claimed ChatGPT or Copilot is "saving hours every day" in software development, I've still yet to see any credible evidence of that.
Aspiring developers: if you're striking out in "coding" interviews of the HackerRank or FizzBuzz variety, take heart in the fact that they rarely have much relevance to actual software development. It doesn't mean you'll necessarily be bad at the job. Chances are, the organisations putting you through these kinds of "coding" interview are not so hot at software development themselves.
It's doubtful generative A.I. will make it possible for non-programmers to build, deploy + maintain software products in the foreseeable future, so the hype about it removing the barrier to entry for founders is probably unwarranted. But you know what *will* remove the barrier to entry? Learning to program. You could start now. It's completely within your control. Or, y'know, keep your fingers crossed for a major breakthrough in artificial intelligence, which is completely beyond your control.
@mnl The only drawback I see is what I've experienced recently of junior devs interacting with those tools, unaware that it's feeding them BS. It's the same problem as copy/paste programming. *Exactly* the same, in fact.
@thirstybear That's the thing: if it doesn't have the information, it can't generate an explanation. It excels at explaining code that's easy to understand. The real selling point is brownie points for comments. Like at university.
@daedalus The person to speak to is Robert C. Martin. It's from Clean Code. I've personally recorded programming sessions (not just mine) and noticed just how much time nothing happens on the screen during "coding" sessions. Various sources (e.g., The Mythical Man- Month, and this one https://blog.ndepend.com/mythical-man-month-10-lines-per-developer-day/) estimate average dev lines of code per day at between 10 - 100. Let's call it 55 LOC/day. Average words in a LOC: ~5. 275 words per day. At 30 wpm, < 10 mins of typing code/day.
We spend roughly 10x as much time reading code as we do writing it. A tool or technique that makes you twice as "productive" at writing code *at best* makes you 5% more productive over all. Making your code easier to understand will have 10x the impact. But that doesn't sell tools or put developers out of work, so you won't be reading about it in Forbes.
Folks debating whether A.I. will be a benefit to the majority of people or simply concentrate even more wealth into even fewer hands should consider the signals the industry's already sending out