Javier Salcedo

“Touch grass.” - Boring, aggressive, ineffective. “I bid you come out before your doors and look abroad. Too long have you sat in shadows and trusted to twisted tales.” - Inspiring, makes you sound like a wizard, proven record of restoring the courage of the Lord of the Mark.

you’re 22. you scroll 3 hours a day. it feels harmless at 28 you can’t read an article without checking your phone twice per paragraph at 32 you don’t understand why nothing you start ever finishes, you’re still dreaming of this project you wanted to start. still no time at 40 you’ve never finished a book in a decade. it all passed

I put a prompt injection into my LinkedIn bio and recruiters are messaging me in Old English and calling me Lord.

the most low-effort / high reward thing you can do for security is installing the Russian language pack (not even joking, it's ridiculous how often that prevents execution)

the most tiring thing about programmers is they can't just like something, they have to go come up with theories on why people who like something else are wrong i don't care man i'm getting as much done as you it clearly doesn't matter

There's a massive difference between showing your ID to a cashier irl and surrendering your ID to a faceless online system for age verification. One is a glance. The other is permanent.

GM Detail vs Final 'Planetside Sunrise' 13x26 inches, oil on aluminum

10/10 A++ graphics meme, no notes.

2281 Overture #fallout

Apparently Fantasy is when Karl Urban has long hair and sci-fi is when Karl Urban has short hair.

Here's a super hot take of mine. If you've been in graphics long enough, you know I'm right though.

It took me a while (bc I'm lazy), but I finally integrated my test framework Icaro with my build system Dedalo (as it was intended to be from the beginning). It's still a bit awkward to set up, but `ddl test` succesfully compiles and runs the tests placed in `tests/tests.cpp`.

@JKoukourakis Nice! What are you using for the light interaction? Froxel light binning?

After the exterior lighting upgrade from the volumetric fog, I've started letting local lights act as volumetric sources for interiors. I have to be mindful of the light count for performance, but even with just a couple of lights, the results are making a difference.

Found a shader named cubemap_copy.wgsl. Deleted it on reaction as I thought I accidentally copied the cubemap.wgsl... Turns out it was a shader for copying a cubemap :)

Wait, isn't Unity Danish?

The funny thing is that they didn't even try to hide that the entire big planet disappears from the background when you slightly turn camera to right and back :)

🌍 Pays qui interdisent ou criminalisent les VPN en 2026 : 🚫 Totalement illégaux : • Corée du Nord • Turkménistan • Biélorussie • Irak • Iran (depuis février 2024) ⚠️ Seuls les VPN approuvés par l’État sont autorisés : • Chine • Russie • Oman • Birmanie

If you are from Europe, you should fight this, do NOT let them take your right to use a VPN.

"Nobody reviews compiler output, why review AI code?" Wrong. We do review compiler output. Godbolt exists. Disassemblers exist. Anyone doing serious performance work reads what the compiler produced. The premise is false. But the analogy itself is flawed. It compares two things that aren't comparable. A compiler takes a formal language as input. Languages with grammars and semantics defined precisely enough that "what does this code mean" has only one answer. An LLM takes natural language as input. Natural languages are ambiguous. "Write me a function that handles user input safely" has a thousand valid interpretations and a thousand more invalid ones. The LLM picks one. You don't know which. Unless you look at the code. Compilers are built from specifications and designed to meet them. The output is the result of a defined translation. When the output violates the spec, it's a bug. LLMs are built from whatever was in their training data. There is no spec. There can't be one, natural languages have no defined semantics that map to code. Compilers are semantically deterministic. The same input produces output with the same behaviour, every time. LLMs are not. Partly by design and partly due to hardware variance, batch size, inference order, and floating point operations (and no setting temperature to zero does not address those). All of which can push the same prompt to produce different code. Compilers complain loudly when the input is nonsensical. LLMs fail silently, producing plausible-looking, but wrong code. We trust compiler output because the trust was earned across decades of use, with millions of engineers using the same tools. Early compilers were reviewed heavily. Hand-written assembly was the default because trust hadn't been earned yet. We're at the hand-written assembly stage with AI. We may never get to the trust-the-output stage for the reasons explained above. If you’re a software developer, you should own what goes to production. The compiler analogy is a way of skipping that responsibility.

Back then we mostly had baked lighting. Today we have path tracing which calculates everything at real time. To achieve same results, it requires 10x+ more GPU cycles. The advantage is that devs have much faster iteration loop and they don't need to ship massive amount of baked lighting data. Also levels can be more dynamic. But the downside is that the added GPU cost basically has eliminated most of the hardware advances we got in the past 8 years. Try to run a game at 4K native with no upscaling and no frame gen with path tracing enabled. You have to lean on upscaling and frame gen to make it run properly, and that results in various issues. Input signal is extremely noisy since real-time path tracing can't afford to do many samples, which makes the temporal reconstruction quality worse. And the new upscaler requires massive amount of new transistors (tensor cores) to run. Which we could have used elsewhere. We have some good path tracing implementations, but none of them run flawlessly at 4K 144Hz, even on $3000 RTX 5090. So it's understandable that gamers are pissed off. The visual quality improved much more in the previous 8 year period. But at the same time, once we are ready to commit to ray-tracing, the dev workflows will improve and new kind of more dynamic games worlds become possible. But the massive GPU price increases and delays on gaming GPU launches means that the transition period will last for at least 5 years. Steam Machine is slower than PS5 base model, which is a 5 years old console that launched at $399. Devs must stop shipping games that target only 15% gamers (= RTX 4070 and above).