hey, asking for a friend can we stop using lines of code for a measurement in productivity? at one point we all agreed on this

Don't fall for this snakeoil. Using LLMs (which are universal function approximators) to approximate GCC, while impressive in a perverse way, is not the same thing as what they want you to think. Ok, LLMs can produce a very poor implementation (of zero commericial value) of a

We spent years and billions of dollars inserting a C compiler into a lossy hash table. Here’s how we spent two weeks getting it back out with a Markov process.

I think that the people who were perceived as fast coders and quick problem solvers by their peers benefit the most from AI assisted tooling and workflows. Their brain was already calibrated to a higher clock rate than others so in fact there might be less mental tax now for them

I wish Google would stop benchmaxxing for long enough to make a usable model. Gemini 3 Pro is as smart as Opus 4.6 but it screws up tool calls as consistently as Grok 3 Mini

LLMs do not think. LLMs do not reason. LLMs have no memory. LLMs have no experience. When an LLM is not producing your output it’s not doing anything. I don’t think this is artificial “intelligence” or any kind of intelligence.

Just had to remind Opus 4.6 that env variables need to be read. Three separate times. And also remind it that it needs a package.json to have packages available. I don't know what they did but this is borderline unusable.

Endless whack-a-mole game. Opus 4.6 is a prime example of a model that over thinks everything, even the simplest changes. It's like the new models are all benchmaxxed and fail at common use cases more and more often.

are people finally waking up

I hate these "coding isn't the hard part" tweets I have been a part of and seen several companies not just struggling with "the right decision" but the culmination of their past technical decisions. AI won't magically make this go away. Lines of Code is still a liability and

Tried to get some PostgreSQL quota from the Microsoft Azure East US regions: MS has zero capacity there and apparently a large backlog of requests.

This is a very apt thing I have noticed a certain malaise i am seeing. it honestly reminds me a lot of 2021 neovim config andies. You can build anything, just some time and some prompts and you got your fully custom piece of tech for exactly what you want!! but wait, here is

Gemini 3.1 Pro is completely unusable: -Endless "Refining Analysis Approach" loops -Failed every coding task I gave it (Opus 4.6 passed) -10x slower than other models I really wonder why Google released this. It's so bad.

lines of code has become a good metric now i so often see simple projects that are noticeably way more LoC than i'd expect great sign of slop

Remember folks: benchmaxxing doesn't produce good models. Gemini has been absolutely awful on most coding tasks already in previous generations and the new models don't seem to be any better.

TIL that Koodiklinikka, Finland's largest software dev community, already has an icon for GitHits 👀

Just think how many vulnerabilities there are in closed source. Most of them never get patched. For OSS, patching is easy due to transparency. And that's why you should use OSS libs.

Which one would you like to see on a SaaS landing page: 1. A video of the app where someone explains how the app works or 2. A replica mockup that you can quickly try out and it replays some real workflow interactively, allows you to do some clicking like in the real app

Whenever you have issues like this, remember that GitHits will help your agents to overcome them. It's not worth spending all your Claude quota to solve problems that can be solved with a few MCP tool calls.