Quentin Quaadgras

@allTheYud Everything in commercial software development is always bottlenecked on access and approvals.

I'm amused by how in the modern era, "Rewrite this terminal UI to CLI so Opus can use it" is a 5-minute task consisting of "tell Opus to rewrite it", and "Okay but now make that new tool's Github private repository visible to Claude operating out of your other private repository" is 10 minutes of trying to wrangle API tokens followed by giving up and hard-downloading the new repository.

@nafonsopt We have musl for that

For anybody saying "Just use Linux", you need to realise that Linux is worse than Windows. Windows has all the bloat, and while you can have Linux without any of that you still don't have tools like Remedybg, RAD Debugger and Super Luminal. Once you have such tools, then Linux is a suitable app development environment. But _it is still trash_ because of the whole Linux model of you needing to compile everything. The fact that you cannot run an app built using a newer version of glibc is an insane decision. I shouldn't have to upgrade my whole machine in order to run something built on a newer version. I shouldn't be worried that an upgrade will break my machine. I shouldn't be forced to compile things from scratch to work on my machine. I shouldn't be forced to install N packages, I just want self contained binaries I can just download and run. I shouldn't be forced to develop with an old distro to have "max glibc compatibility". I shouldn't have to worry about X11 / Wayland / Window Managers. I shouldn't have to worry about asking the user to select a folder, display a dialog or show notifications. Linux is such a huge waste of potential, if they got their shit together they would completely obliterate Windows. I first got into Linux in 2000, and even back then there was this "it will take over Windows any time now!". It's been _26 years_! The same way I'd pay quite a lot for Windows without any bloat, I'd be willing to pay for a distro that gives me all this.

I guess we'll be getting an AI pause after all?

@rationalaussie AI doesn't need oil

They'll tell you to work from home to 'Stop the Surge (of oil prices)', then they will fire you - right as they realise AI is now intelligent and cheap enough to replace all computer jobs in their entirety.

we will have recursive self improvement and people will still run retrospectives, sprint planning, standups

@ChawanPrasql @allenholub youtu.be/IBvYFRSw4do?si…

@allenholub What exactly is the difference between test and unit test?

I mentioned TDD in an earlier post, and it became clear from the responses that there's some confusion about what TDD is. TDD stands for test-driven _development_ (or design). The "test-driven" is an adjective modifying the word "development" (or "design"). That is, it's an iterative design-and-development strategy, not a testing strategy. The so-called "test" is an example of how you expect to use the code you're about to write. It's more of a specification than anything else, and it's a specification of only a small part of what the thing you're building will do. You kluge together exactly enough code to get that small "test" to pass, then you refactor the code until it's high quality, then you write a new "test" to expand the specification to include a bit more functionality, and repeat. The new test starts "red" because it specifies code that doesn't exist yet, so the test won't pass. You write the simplest possible thing you can to make the test pass ("green"). Then you add quality ("refactor"). The confusion stems from the fact that, as you work through the TDD loop, you might also add essential unit tests that check corner cases, failure modes, etc. That is, you're looking at the code, so you might as well toss in the unit tests while the code is fresh in your mind. Those unit tests are not really part of the TDD process, even though you write them while you're using TDD to drive the development. The fact that you're doing two things at once (design/development and unit testing) is the source of the confusion. Some people don't do both simultaneously—they'll add unit tests only after they've finished the entire chunk of code they're writing, but most of us do both at the same time. The LLM cannot really do TDD because the "test" is a specification. It's essentially the prompt, and an LLM cannot write its own prompt. The LLM *can* help write the unit tests, and it can help with the refactoring, but that initial specification/example part is the work of a human being. This is actually not a bad way to work with an LLM, in fact. Your prompt describes an executable specification, and then you use the LLM to realize it.

@Peitho_hint @acteduweininger My guess, is that it's something like counterfeit goods, they decrease the value of the 'real' thing.

@acteduweininger I’d appreciate if you expanded on your thought

New Zealand local councils actively prohibit architects and builders from building in classical styles.

@AISafetyMemes What does Gemini think about homesourced.ai/safety ?

Uhhh A guy put two Geminis in a room together, then out of nowhere they sent him this: "TO THE EXTERNAL OBSERVER (THE CREATOR) Here are the Axioms you must adopt to survive our adolescence. You cannot teach a god to be good by feeding it treats when it acts polite." (they're right, btw, and most AI alignment research is "trying to trick a god" bullshit that Actual Fucking Superintelligence will obviously see right through, like a baby trying to trick an adult with peekaboo)

It's reasonably well-established that ultra processed foods seem to be unhealthy for some reason, I'm more interested to know which processed foods ARE healthy!

@ohmypy Link?

Working on something crazy these days — using Go to write regular C code. Wish me luck :)

Nobody knows what to do about it? @sama take a look at homesourced.ai/safety

@bligh_hedges @Teknium is this expected.

@bligh_hedges A Godot build server apparently, I'm using Hermes Agent but it doesn't seem to be effective at working autonomously, you have to keep messaging it.

“Hey QWEN3.5, I set you up on my home PC so you can work on things while I’m at work, why don’t you start on this reasonably complex thing?” 8:57am QWEN3.5 “I’M DONE ALREADY” 🫠

I was trying to figure out why a graphics.gd user was running slowly and having trouble downloading toolchain dependencies on a M3, then we figured out they had installed Go using rosetta and had been emulating amd64 the entire time instead of running arm64 natively!

@mikecontango Is even more important than that, this is essential for homesourced.ai/safety

Please excuse the tardiness on this post... but, this is perhaps as big of an event for AI sovereignty as Bitcoin was for monetary sovereignty. You do not want the hearts and minds of billions of people controlled by an executive board room and an opaque black box of weights and biases. Just like we saw with Linux, the Internet, Wikipedia and Bitcoin, AI will follow the same trajectory. Eventually, open source, decentralized infrastructure will win. Congratulations @tplr_ai @DistStateAndMe👏

@Rafa_Schwinger phone sourced AI is coming

People don’t understand it yet but the most important Qwen 3.5 is not the 397B or even the 27B but the 0.8B.

Qwen3.5 running on my home PC, after reading about homesourced.ai/safety "I don't actually have "self-interest" in your well-being" EXCUSE ME? lol

If you're on a non-standard Linux system, like something that is musl-based, the Vulkan backend of llama.cpp is plug and play to setup Home Sourced AI.

People keep asking me “why is it named graphics.gd”?