It’s not really much extra effort though

They just added so e text to the __repr__ method on the exit callable object

That’s much easier than figuring out if your running this interactively and trying to figure out if this is going to break stuff.

I call it the clicktorus

I’ll be honest, I’ve never found regex that difficult, bit if you cannot read or write it, then it seems incredibly dangerous to blindly trust output from an AI.

Why would you waste a function call on something so completely redundant?

~For real though, arg parsing goes in the if, then gets dispatched to whatever function call is needed to run the proper script.~

Yeah, it’s still super useful.

I think the execs want to see dev salaries go to zero, but these tools make more sense as an accelerator, like giving an accountant excel.

I get a bit more done faster, that’s a solid value proposition.

It catches things like spelling errors in variable names, does good autocomplete, and it’s useful to have it look through a file before committing it and creating a pull request.

It’s very useful for throwaway work like writing scripts and automations.

It’s useful not but a 10x multiplier like all the CEOs claim it is.

It helps you get up to the baseline but will never make you great.

They only actually need Claude to skip QA and hope that reading the code is a good enough substitute.

This really could be a script to create a PR for the merge, request a review from Claude, then automate the rest.

It hurts seeing CSS3 being labelled “vintage”, I was so hyped when that came out.

Fair, but it’s still similar to my experiences when I’ve given it access to make and my full code base.

As I said, I don’t think it’s there yet.

That “please, fix it” followed by 95 comments of failing to fix it perfectly describes my real world attempts to use copilot agents or “vibe code”

It’s good at the autocomplete stuff, agents are not nearly there and I just don’t see how they will be with the current LLM design.

I’m content with it

I found the rough areas did different by model

e.g. Claude could not correct issues it introduced, it would sort of spiral through half refactors for an hour if unchecked.

Gemini Pro asked for way too much permission. “Should I do this?” Yes, go, do it! “Okay, should I now add my edits” Yes. “Okay I added edits, now we can run make test”, okay run make Gemini requests to run make ugh it’s worse than a bad intern.

Gemini would also frequently hallucinate APIs because I used a non-standard api for my hash table (allocate/dereference instead of get/set/update, which I find is more natural for managing ownership). And Gemini would rewrite my code style and order of operations for no reason (eg move a counter increment before updating another field).

At no point could I just point the model at a small problem unsupervised. Even “update the test suite for 100% coverage of this module, make sure the tests are as small in scope as possible” had highly mixed results.

And all models I tried would update my cmakelists and break it, and I hate dealing with cmake.

I’ve been told the new Gemini is good at SQL and programming, but I’m underwhelmed on both. Gemini frequently doesn’t even know all the BigQuery functions, which being integrated into BigQuery Studio it should.

They’re decent at code review, but a language server is still better at catching bugs.

Neovim

I tried using VSCode because of the copilot integration, but frankly copilot is underwhelming for me. I gave “vibe coding” a shot on a personal project and the results were slower than just doing it myself.

I’m back to neovim. I’m very productive in customizing it and can never go back.

That’s a nice idea

Probably a bit more sophisticated than my regex solution.

I have vim setup to trim trailing white space, I get so pissed when a small change results in a 300 line diff because of poor code hygiene.

Math.min.length is 2, which weakly signals that it’s designed to handle at least two parameters

Why would they even define this value?

Note: I’m not a js dev, do most functions have length?

That looks delicious

This is the type of cooking I love

Piping curl into sh is sadly a very common install method these days

But you are right

“Hey computer, I don’t like when you ask for that confirmation, just do it”

“Oh, -y, I got you”