|
Large Language Models and Simple, Stupid Bugsquote:With the advent of powerful neural language models, AI-based systems to assist developers in coding tasks are becoming widely available; Copilot is one such system. Copilot uses Codex, a large language model (LLM), to complete code conditioned on a preceding "prompt". Codex, however, is trained on public GitHub repositories, viz., on code that may include bugs and vulnerabilities. Previous studies [1], [2] show Codex reproduces vulnerabilities seen in training. In this study, we examine how prone Codex is to generate an interesting bug category, single statement bugs, commonly referred to as simple, stupid bugs or SStuBs in the MSR community. We find that Codex and similar LLMs do help avoid some SStuBs, but do produce known, verbatim SStuBs as much as 2x as likely than known, verbatim correct code. We explore the consequences of the Codex generated SStuBs and propose avoidance strategies that suggest the possibility of reducing the production of known, verbatim SStubs, and increase the possibility of producing known, verbatim fixes.
|
#
¿
Apr 3, 2023 18:24
|
|
|
|
| # ¿ May 10, 2024 05:12 |
|
|
Judge Schnoopy posted:I mean, if I just copy and pasted chatgpt code I would get eviscerated at code review. That depends on what code safety means. If you're just writing a CRUD app, inspection is good enough, but if you're doing any kind of non-trivial multithreading it can be quite a bit tougher to figure out whether your code works right.
|
|
#
¿
May 25, 2023 23:52
|
|
|
Just tell them 4 is the xth Fibonacci number where x ≈ 4.549.
|
|
#
¿
Jun 7, 2023 04:03
|
|
|
AmbientParadox posted:whats the difference between a v&v engineer and a test engineer? Industry, mostly. Verification and validation are specific testing-related activities defined in ISO 9001, so if you see a job posting for a v&v engineer you can be pretty sure it's in an industry that cares about that particular standard.
|
|
#
¿
Jul 19, 2023 15:00
|
|
|
A few jobs ago one of my colleagues gave a phone screen where he asked for the complexity of a piece of code, and the candidate told him it was complex because it had a lot of loops.
|
|
#
¿
Aug 22, 2023 04:03
|
|
|
Carbon dioxide posted:I don't see the problem? Maybe they didn't describe it properly, but Computational Complexity is literally the number of branching paths your code can take, and any non-infinite loop counts as a branch. You're thinking of cyclomatic complexity, and that's slightly more complicated than what you describe. It's also not what any interviewer is looking for when they talk about complexity.
|
|
#
¿
Aug 22, 2023 14:26
|
|
|
Changing names here, but we have two guys named Joe, two guys named Rob, two guys named Sam and two guys named Seth. We had two guys named Matt and another Rob up until about a month ago. Our stakeholders include two guys named Rob, two guys named Seth, a guy named Matt, and a guy named Joe. I'm probably forgetting a few.
|
|
#
¿
Sep 21, 2023 18:04
|
|
|
Macichne Leainig posted:Well gently caress, I said too much and now I'm a VP reporting directly to the CEO. 
|
|
#
¿
Oct 25, 2023 20:35
|
|
|
Gin_Rummy posted:Normally I’d agree, but he doesn’t seem too keen on receiving feedback. I tried to apply logic reasoning against what he considered a blocker (also, it wasn’t even a blocker) so that we could keep progressing, but he took it personally and then just disappeared for the rest of the day. He seems to have a complex where if you don’t agree with him, he will take his ball and just go home. Document all of your interactions with him so you have a paper trail.
|
|
#
¿
Oct 31, 2023 16:57
|
|
|
Falcon2001 posted:Yeah, adding onto this. It also speeds up onboarding folks who might be familiar with the language, but aren't familiar with the specific toolset you're using (or may have been working somewhere with weird environment setups before). And you can get them to update it to what they actually need to do as part of their onboarding.
|
|
#
¿
Nov 12, 2023 15:43
|
|
|
The career path thread over in BFC is probably the best place to ask.
|
|
#
¿
Dec 9, 2023 05:35
|
|
|
Do Users Write More Insecure Code with AI Assistants?quote:We conduct the first large-scale user study examining how users interact with an AI Code assistant to solve a variety of security related tasks across different programming languages. Overall, we find that participants who had access to an AI assistant based on OpenAI's codex-davinci-002 model wrote significantly less secure code than those without access. Additionally, participants with access to an AI assistant were more likely to believe they wrote secure code than those without access to the AI assistant. Furthermore, we find that participants who trusted the AI less and engaged more with the language and format of their prompts (e.g. re-phrasing, adjusting temperature) provided code with fewer security vulnerabilities. Finally, in order to better inform the design of future AI-based Code assistants, we provide an in-depth analysis of participants' language and interaction behavior, as well as release our user interface as an instrument to conduct similar studies in the future.
|
|
#
¿
Mar 12, 2024 22:43
|
|
|
Love Stole the Day posted:How do we monetize this? Like so: code:
|
|
#
¿
Mar 26, 2024 20:46
|
|
|
|
| # ¿ May 10, 2024 05:12 |
|
|
I never really appreciated what product managers do until I worked on a team where the business had direct access to the engineers. What a miserable way to run things.
|
|
#
¿
May 1, 2024 18:30
|
|