I’ve reviewed GitHub copilot code and failed it. Despite a lot of people in my org being all-in on ai, I remain sceptical of quality (and ethics) for many usecases
Ugh, imagine having weekly 1:1 meetings with all your AI direct reports
640K ought to be enough for anyone
- Bill Gates (allegedly)
Then I’m no future dev. 🖕🤖🖕✌️
I think the title mischaracterises the content a bit. IMO a better summary would “in the future software development will be with an AI partner”. The author says developers are saying (which developers, I don’t know) that 90% of code in the future will be written by AI, but CompSci fundamentals will remain critical to understand what the AI did. However CompSci will only be one fundamental as AI collaboration and software architecture will become the others.
The author also believes that CompSci courses will have to change in a big way to enable this new way of development. Memorizing syntax and hand-written code that’s graded on syntax or code style are out of date (surprise, they have been for at least 2 decades). New testing will be necessary.
All in all, I do believe this is a change we can’t stop. Whether it will be good, is yet to be seen. There will be bad devs who don’t understand what the AI does and just copy paste and here will be those that can harness AI effectively - if it can be harnessed effectively 🤷♂️
The author claims to be optimistically realistic om the subject, but I would claim they are more optimistic than realistic and that’s of course because they have to be. Their paycheck depends on the AI gamble.
