Be advised: this thread is not for serious in-depth discussion of weighty topics (we have a link for that), this thread is not for anything Culture War related. This thread is for Fun. You got jokes? Share 'em. You got silly questions? Ask 'em.
- 115
- 1
Jump in the discussion.
No email address required.
Notes -
Well, it's happened. The pressures of our workload have driven a push to use AI to help with some of the bitchwork of coding. There is going to be a push to use AI to generate unit tests.
The toolchain involves OpenCode, so I figured I'd install it locally to get familiar with it before I start burning GPU time at work. Also, for reasons, we aren't allowed to use the GPUs. So last night I installed Ollama, OpenCode and the gemma4:e4b model on my humble RTX 4070 Super with 12 GB of VRAM. I tried to have it do the simplest of tasks. Create a Hello World project in dotnet10, and write a single unit test to verify it's output.
The first thing that happened was it created a new project. New projects actually begin with hello world output already. It then added a second hello world output. This poisoned the well, as now the AI was horribly confused about why there were two hello world lines. It never fully recovered. The project was generated without it's int main format which I prefer, so I tried to have it restructure the project to use that. After several missteps because it couldn't get over the fact that there were duplicate hello world statements, it finally figured it out.
Next came the unit tests. It created a unit test project, but then didn't actually populate the tests or link the projects. Then it wanted to refactor hello world, and pull in all sorts of abstraction frameworks, so it could test the output without redirecting stdout. I told it forget all that and redirect stdout. It had already done half the refactoring in a state that could not compile, and then never undid it, and then the whole project was totally broken and it couldn't figure out how to fix it.
I remind you, this was a "Hello world" and a single unit test. I told the AI it fucked everything up, it asked what it could do to fix it, and I told it that it could shutdown. It did. I think.
I know a lot of people reading this are AI evangelist. Where did I go wrong? What the fuck do people see in this shit?
Your results are worse than mine, but I'm in a similar boat. The results that the boosters are seeing are not what I am seeing in my professional life.
I can approximate what the boosters are seeing in personal projects, if I keep the scope small, and I use python, and the project doesn't involve much state management, and I write a spec that is roughly the same size as the codebase that would exist if I were to write it by hand.
This kind of tracks with the handful of AI enthusiasts I know in real life, too. The first category is professional programmers. When I ask what problems they solve with it, they tell me they're using AI to write an AI agent orchestration framework to better run their AI. When I ask what they'll use that tool to do, I'm told to shut up, because AI is the future and you can't be left behind. The second category is non programmers who ask Claude to process a CSV file, and this blows their dick clean off. The idea that you can process a CSV file and you don't even need Excel to do it is a mind-altering experience for them. The fact that Claude is doing a tool call that runs ten lines of python is irrelevant. This is the future. They'll keep rebuilding that same deterministic python script from an elaborate prompt every single month, because holy mother of cyber-jesus, this saves them half an hour in Excel every month.
Everyone else I know is either possessed of a terrifying antipathy, or can't muster up anything more enthusiastic than "it's ok, I guess".
On an unrelated note - do you know why .net projects like to keep their unit tests in a segregated project? Most other languages that I've used have a convention where the unit tests are in a separate folder in the same project.
I donno, why do most other languages comingle their production code with their unit tests? I can't speak for any official reason, but I know managing large dotnet solutions, testing projects need to be flagged with true in it's project file or some shit like that. Otherwise test runners don't find it. They usually also have a bunch of nuget packages pulled in, like xunit or MSTest that your production project won't need. They'll probably have some resources embedded in them to facilitate testing which you almost certainly do not want shipped in production.
Man, I was checking the dependencies xunit3's nuget pulls in now. Shit looks like a maven package.
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link