Hacker News: diegoperini

New comment by diegoperini in "Gemini API File Search is now multimodal"

diegoperini — Sun, 10 May 2026 10:21:15 +0000

I have the opposite experience where Gemini (even the flash models) has the only useful model for my reverse engineering related use case. My hunch is Google utilizes its free access to entire Google search indices to train itself from niche non-English speaking community websites, much frequently and in a "relevant" manner, which in the end gives these models the most up to date info for this particular kind of work. Every other model is just either 10 years outdated with their answers or simply hallucinates like waaaay crazy.

New comment by diegoperini in "Tell HN: An app is silently installing itself on my iPhone every day"

diegoperini — Sun, 26 Apr 2026 05:19:42 +0000

If I am not mistaken, it's even shown in the marketing materials to build suspense.

New comment by diegoperini in "All 12 moonwalkers had "lunar hay fever" from dust smelling like gunpowder (2018)"

diegoperini — Sat, 18 Apr 2026 16:25:29 +0000

> Water blobs vs. fireballs. Pretty sure there's a nice videogame idea hiding in there somewhere.

I like the way you think. :)

New comment by diegoperini in "When if is just a function"

diegoperini — Sat, 18 Oct 2025 16:17:28 +0000

That analogy may not be suitable for this case because value proposition between the aesthetics vs the function is different for visual art projects compared to software. There is also the maintainability factor where most aged software (especially the closed source ones in private sector) change maintainers every few years. Old maintainers most often lose access to the source code and become unreachable after leaving their job.

New comment by diegoperini in "Claude Haiku 4.5"

diegoperini — Thu, 16 Oct 2025 14:22:16 +0000

"Protection against malicious use" isn't as cool as "model welfare". I'm renaming my authentication function to "examineCrest()".

New comment by diegoperini in "Show HN: I made an esoteric programming language that's read like a spellbook"

diegoperini — Sun, 12 Oct 2025 16:50:04 +0000

I recommend replacing "hi chat" in the test spell with "Mortal plane, I greet thee".

New comment by diegoperini in "Erlang ARM32 JIT is born"

diegoperini — Tue, 07 Oct 2025 16:19:26 +0000

> I'm not quite sure why you'd want to run Erlang on it, but the hardware exists.

Erlang is invented before IoT was a thing to facilitate distributed computing for telecommunication in a highly reliable manner. It makes perfect sense to adapt it for driving fleets of cheap IoT devices.

New comment by diegoperini in "The G in GPU is for Graphics damnit"

diegoperini — Mon, 06 Oct 2025 03:43:59 +0000

> And here are my clearly unimpressed “friends” >:(

These friends don't get it!

New comment by diegoperini in "Traefik's 10-year anniversary"

diegoperini — Fri, 26 Sep 2025 15:36:36 +0000

Low resource footprint, written in Go, embed-able in any Go project as a library, compiles to mobile with little to no modification, supports config change without restart, has plugin API.

These were the reasons why we used it in my previous job.

New comment by diegoperini in "We Politely Insist: Your LLM Must Learn the Persian Art of Taarof"

diegoperini — Mon, 22 Sep 2025 11:25:16 +0000

> it has a truckload of data

It would be "artificial" only if LLMs performed badly despite having an equal amount of data containing examples of eastern customs in its training set. Even that's arguable since we don't (didn't) have the benchmarks for this particular case before.

New comment by diegoperini in "We Politely Insist: Your LLM Must Learn the Persian Art of Taarof"

diegoperini — Mon, 22 Sep 2025 10:43:11 +0000

It's no different than GPT answering a prompt with "That's a wonderful idea!", except it's in a different language than English. It's a good thing if LLMs can do this in every language and for any culture with no compromise.

New comment by diegoperini in "Open source can't coordinate?"

diegoperini — Fri, 20 Jun 2025 04:27:54 +0000

> yet the end result was complete shit

Could you elaborate why? It looks like a useful protocol.

New comment by diegoperini in "Fan Service"

diegoperini — Sun, 11 May 2025 09:06:42 +0000

> Linux separates things such that I was looking at C files in drivers/platform/x86 and header files in include/linux/platform_data/x86. And the ACPI code lives other places as well. It’s all very orderly, but at times it felt like navigating a grocery store that arranges products in alphabetical order. Logical, but not exactly cozy.

So beautifully put!

New comment by diegoperini in "SeedLM: Compressing LLM Weights into Seeds of Pseudo-Random Generators"

diegoperini — Sun, 06 Apr 2025 22:39:42 +0000

You get to choose your own, more efficient "PI" for your model. Still, it's a valid question.

New comment by diegoperini in "But good sir, what is electricity?"

diegoperini — Mon, 24 Feb 2025 13:30:30 +0000

I refuse to believe neutrons can be any color other than white.

New comment by diegoperini in "Phind 2: AI search with visual answers and multi-step reasoning"

diegoperini — Thu, 13 Feb 2025 20:50:16 +0000

Phind does one thing and it does it really well. I use it because it has a real, positive impact on my growth. I love the fact that it is not Jarvis, but just a really helpful library assistant. Please don't lose the focus! Stay awesome.

New comment by diegoperini in "Don't "optimize" conditional moves in shaders with mix()+step()"

diegoperini — Sun, 09 Feb 2025 15:23:38 +0000

Double "AFAIK" makes me trust you more :)

I think this is very accurate. The exception is probably those block buster games. Those probably get direct consultancy from NVIDIA during the development to make them NVIDIA-ready from day 1.

New comment by diegoperini in "Emergence of collective oscillations in human crowds"

diegoperini — Sun, 09 Feb 2025 14:49:55 +0000

Beautiful. Thank you.

New comment by diegoperini in "Emergence of a second law of thermodynamics in isolated quantum systems"

diegoperini — Mon, 03 Feb 2025 11:50:15 +0000

State transitions are probabilistic and operators have complex coefficients.

New comment by diegoperini in "--libcurl"

diegoperini — Mon, 29 Jan 2024 16:31:34 +0000

True. Luckily it's open source. You can do `npx curlconverter --language go example.com` behind a firewall after downloading the npm module.