It's popular because they have the best models and they are burning obscene amounts of money handing out tokens via subscriptions (for now) at a huge discount compared to what the API costs are.
Claude Code itself is incredibly buggy and as we have seen the codebase is a complete mess of slop.
Bizarre to say that. When I have it perform work on a bespoke code base on a niche videogame, in a less commonly used language, is that still "regurgitating stuff"?
No, it is impossible for it to have seen this combination of things.
It routinely produces, suggests, and correctly implements novel things that had not existed.
You can see this yourself by learning how LLMs work, or anecdotally using these tools.
LLMs are terrible at generating code for “less commonly used languages”. They require LOTS of data for high accuracy.
I describe it this way: they are good at interpolating from what data they were trained on, but terrible at extrapolating. I agree with the parent that the LLM-generated content isn’t novel, it’s just a rehash of two things it was trained on.
I have wasted quite a number of hours trying to use LLMs to write things for less common languages. Sure they can one-shot some impressive stuff in C#, Python, and JavaScript… but try working in Object Pascal: it’s non-obvious hallucination after non-obvious hallucination, presented confidently enough to make it difficult to see it’s complete garbage, so you waste a ton of time trying to polish a turd.
Creating a new paradigm is a problem with a lot more groundwork laid that working in an existing little-known paradigm. One is creating patterns which only have to be good to be correct. The other has to be correct to be good. They are completely different problems.
That is simply not true. The naive “glorified auto-complete / stochastic parrot” argument may have some merit when applied to generic pre-trained models, which only learn from unsupervised next-token prediction. But the post training through reinforcement learning the frontier models undergo is very sophisticated and they genuinely learn to do novel things that are purely the work of the model being trained (and the work of the GPUs they burn along the way of course).
I would argue it couldn't be more different. I can dive into the source code of any library, inspect it. I can assess how reliable a library is and how popular. Bugs aside, libraries are deterministic. I don't see why this parallel keeps getting made over and over again.
I can dive into the source code of LLM generated code too. Indeed it is better because you have tools to document it better than a library that you use.
One path is that you could try to transitions within your current org. This should be particularly easy in startups, which is where you say your experience is, as startups have a lot less rigidity in roles/responsibilities and you could contribute to infrastructure efforts to build up context/knowledge.
From there, you can leverage that into an "official" infra role.
I do wish they'd focus on closing the gap to Jetbrains by implementing the QOL features that are missing. I understand they have to do what VC wants to see, but this agentic stuff is so tiring.
I give them a try about twice a year. I write a lot of Rust which should be squarely in their wheelhouse.
This last time I was pleasantly surprised to find they mostly fixed their SSH remote editing support. But then it started truncating rustc inline error messages and I couldn’t figure out how to view the whole thing easily. When you’re just trying to get something done little bits like this can add up quickly. Punted back to Cursor for now.
I don't like the way remote editing works with plugins. IIRC, the remote agent pulls the plugins from the connecting client. I get why it's done like that, but I'd way rather have it go the opposite direction.
I want a setup where I can have an immutable devcontainer with local copies of everything I need to develop 100% offline; dependencies, tools, etc.. Having my local editor pull plugins from a devcontainer for the project seems to make more sense to me.
I didn't dig in too much. Maybe there's a way to make it work somehow.
Yeah, Zed is why next year I might not renew with JetBrains at all. They hiked their price for AI stuff I didn't really ask for, especially when Claude Code does everything I want and need.
When will this boring line keep getting posted. I literally use my tesla every time with FSD and browse twitter while it drives. At most I have to glance up every 3 minutes to avoid the alert, which I'm guessing they are obligated to by law.
I'm in the same boat. Much like Elon has also promised, I'm also using my Tesla as a robo taxi when I'm not in it myself, having it earn passive income for me has been a massive success.
I have unfortunately found myself doing stuff like this too, although maybe not as egregious.
I think part of the problem is that our brains are wired to look for the path of least resistance, and so shoving everything into an LLM prompt becomes an easy escape hatch. I'm trying to combat this myself, but finding it not trivial, to be honest. All these tools are kind of just making me lazier week over week.
There’s some kind of new failure mode here. People seem to determine a tool’s applicability for a task by whether its interface allows for their request to be entered. An open ended natural language input field lets people enter any request, regardless of the underlying tool’s suitability.
Claude Code itself is incredibly buggy and as we have seen the codebase is a complete mess of slop.
reply