> This seems like it's fixing the symptom rather than the underlying issue?
This is also my experience when you haven't setup a proper system prompt to address this for everything an LLM does. Funniest PRs are the ones that "resolves" test failures by removing/commenting out the test cases, or change the assertions. Googles and Microsofts models seems more likely to do this than OpenAIs and Anthropics models, I wonder if there is some difference in their internal processes that are leaking through here?
The same PR as the quote above continues with 3 more messages before the human seemingly gives up:
> please take a look
> Your new tests aren't being run because the new file wasn't added to the csproj
> Your added tests are failing.
I can't imagine how the people who have to deal with this are feeling. It's like you have a junior developer except they don't even read what you're telling them, and have 0 agency to understand what they're actually doing.
Another PR: https://github.com/dotnet/runtime/pull/115732/files
How are people reviewing that? 90% of the page height is taken up by "Check failure", can hardly see the code/diff at all. And as a cherry on top, the unit test has a comment that say "Test expressions mentioned in the issue". This whole thing would be fucking hilarious if I didn't feel so bad for the humans who are on the other side of this.
That comparison is awful. I work with quite a few Junior developers and they can be competent. Certainly don't make the silly mistakes that LLMs do, don't need nearly as much handholding, and tend to learn pretty quickly so I don't have to keep repeating myself.
LLMs are decent code assistants when used with care, and can do a lot of heavy lifting, they certainly speed me up when I have a clear picture of what I want to do, and they are good to bounce off ideas when I am planning for something. That said, I really don't see how it could meaningfully replace an intern however, much less an actual developer.
Nice to see that Microsoft has automated that, failure will be cheaper now.
This level of smugness is why outsourcing still continues to exist. The kind of things you talk about were rare. And were mostly exaggerated to create anti-outsourcing narrative. None of that led to outsourcing actually going away simply because people are actually getting good work done.
Bad quality things are cheap != All cheap things are bad.
Same will work with AI too, while people continue to crap on AI, things will only improve, people will be more productive with AI, get more and bigger things done for cheaper and better. This is just inevitable given how things are going now.
>>There's a PM who takes your task and gives it to a "developer" who potentially has never actually written a line of code, but maybe they've built a WordPress site by pointing and clicking in Elementor or something.
In the peak of outsourcing wave. Both the call center people and IT services people had internal training and graduation standards that were quite brutal and mad attrition rates.
Exams often went along the lines of having to write whole ass projects without internet help in hours. Theory exams that had like -2 marks on getting things wrong. Dozens of exams, projects, coding exams, on-floor internships, project interviews.
>>After dozens of hours billed you will, in fact, get code where the new file wasn't added to the csproj or something like that, and when you point it out, they will bill another 20 hours, and send you a new copy of the project, where the test always fails. It's exactly like this.
Most IT services billing had pivoted away from hourly billing, to fixed time and material in the 2000s itself.
>>It's exactly like this.
Very much like outsourcing. AI is here to stay man. Deal with it. Its not going anywhere. For like $20 a month, companies will have same capability as a full time junior dev.
This is NOT going away. Its here to stay. And will only get better with time.
I used upwork (when it was elance) quite a lot in a startup I was running at the time, so I have direct experience of this and its _not_ a lie or "mostly exaggerated", it was a very real effect.
The trick was always to weed out these types by posting a very limited job for a cheap amount and accepting around five or more bids from broad prices in order to review the developers. Whoever is actually competent then gets the work you actually wanted done in the first place. I found plenty of competant devs at competitive prices this way but some of the submissions I got from the others were laughable. But you just accept the work, pay them their small fee, and never speak to them again.