Not a new article https://www.mckinsey.com/capabilities/mckinsey-digital/our-insights/the-economic-potential-of-generative-ai-the-next-productivity-frontier#key-insights It has an interesting (and scary) perspective of SWE as a career. "According to our analysis, the direct impact of AI on the productivity of software engineering could range from 20 to 45 percent of current annual spending on the function. This value would arise primarily from reducing time spent on certain activities, such as generating initial code drafts, code correction and refactoring, root-cause analysis, and generating new system designs. By accelerating the coding process, generative AI could push the skill sets and capabilities needed in software engineering toward code and architecture design. One study found that software developers using Microsoft’s GitHub Copilot completed tasks 56 percent faster than those not using the tool."
Article written by people who a) Don’t have expertise in Gen AI (to be able to speak critically of its potential) b) Are not software engineers or haven’t held positions requiring non-trivial skill-set
Exactly, these consultants are clueless kids
They are overestimating how much time is actually spent on writing code?
This one I agree with
Spreadsheet warriors have no idea. We’ll automate the automatable and manually do the rest of the work, as always. SWE FTW!
I hope this is true. I would be rather working with AI coworker than working with some bootcamper. At least I don’t need to teach them basic CS shit.
Yeah. I don't think it works like that. They give the job to the cheaper of you two and have an LLM writing the code
And there are probably bootcamper who would also rather work with AI than Indeed.
Didn't these same idiots come out claiming they have finally solved how to measure software developer productivity?
Such articles are ghostwritten by 22-25 year old kids at 10pm on a Sunday with the same level of bullshittery as a high school essay. There is no value to them whatsoever.
And the sad part is that's what Sundar reads on Monday morning.
Not Sundar but the CFO. Then Sundar has to argue with her.
Mostly laughable analysis. It will help a lot with initial code drafts, quick one liners that you would normally have to consult documentation or stack overflow for, possibly architectural suggestions, but not much past that This will have minimal impact on the industry, and the impact it does have will just cause engineers to be more efficient and thus more desired. Please see: Jevons paradox
56% faster with copilot seems really dubious, unless it was compared with using just notepad... Or maybe only extremely simple tasks.
I wish there's a way to actually profit from taking a bet against these snakeoil salesmen, rather than passively waiting it out. Unfortunately, there aren't many mechanisms to do that, because it's not very transparent who end up buying into their crap. Right now, they just make noise, make a mess, and move on to the next fad, and all we can do is to watch and hope companies don't fall for their bs.
First of all: Duh Second of all: Good Third of all: Every productivity enhancement in history has been met with not reduced spend, but increased output. Don’t be an idiot and assume companies will try to just do the same thing for cheaper and with fewer people. They will spend the capital they have to do and achieve more. Now if your value to the world is extremely inflexible and/or you’re barely useful as-is, yeah the shot to the arm here is going to hurt. But for real engineers—it’s less time fucking with Bash syntax and making spending precious life minutes writing an RCA
Fourth of all: LOL McKinsey
Real engineer?