Hacker News: bigyabai

New comment by bigyabai in "[dead]"

bigyabai — Fri, 17 Apr 2026 00:39:50 +0000

> When I don’t have the skill to build a good app, what is my best bet. To hope AI does it for me or to grind and learn?

You can do both. I can't design a UI to save my life, but I know how GTK's backend works and I can pick the widgets I like from a gallery: https://gnome.pages.gitlab.gnome.org/libadwaita/doc/main/wid...

If I can describe the architecture/stack that I want for an app, most AI models can build the desktop app for me. You have to remain conscious of the tools you're using, but AI can produce great boilerplate for this type of rote programming.

> Like have you seen cracked coders working, they are like fast apm coders and worlds more accurate then AI.

I've worked with 10x engineers before, most of them were not Carmack-level geniuses. They simply understood their stack and stayed laser-focused on the goal. 10xers are impressive because they're specialists, and most of them go back to the 1-2x level if you force them to use technology they don't understand.

New comment by bigyabai in "Qwen3.6-35B-A3B: Agentic coding power, now open to all"

bigyabai — Fri, 17 Apr 2026 00:20:29 +0000

UMA removes the PCIe bottleneck and replaces it with a memory controller + bandwidth bottleneck. For most high-performance GPUs, that would be a direct downgrade.

New comment by bigyabai in "Brave Origin"

bigyabai — Fri, 17 Apr 2026 00:07:07 +0000

> Brave Origin is a minimalist version of Brave that allows users to disable the revenue-generating features that otherwise support Brave as a business.

Oh cool, we used to call these a "web browser" when I was growing up.

New comment by bigyabai in "[dead]"

bigyabai — Thu, 16 Apr 2026 23:50:42 +0000

> Strauva (Strava for Claude Code yes).

I feel like if you're using this, you're subconsciously focused on gaming metrics instead of getting stuff done. You'd have the same issue scoring #1 on a SLOC tracker - you'd be the king of a slop empire.

Given that this is the second time you've written a piece like this, maybe you can offer us some insight into your workflow? What stack are you using, what's your biggest challenge?

New comment by bigyabai in "Qwen3.6-35B-A3B on my laptop drew me a better pelican than Claude Opus 4.7"

bigyabai — Thu, 16 Apr 2026 21:59:01 +0000

Let's not oversell Opus' output. The Qwen flamingo is flawed but could be easily fixed with 1-2 prompts if you're really upset with it. The Opus SVG is not any better than something that I could make in Inkscape with 3 minutes and sufficient motivation. Calling Opus' flamingo "programmer art" would be an insult to programmers.

New comment by bigyabai in "Five men control AI. Who should control them?"

bigyabai — Thu, 16 Apr 2026 19:50:13 +0000

The law is only relevant insofar as it's enforced. In America, that's a tossup.

New comment by bigyabai in "Google, Pentagon discuss classified AI deal, the Information reports"

bigyabai — Thu, 16 Apr 2026 18:45:13 +0000

The American public only follows pop culture sentiments like "AI slop" and "they made $THING woke", we can't expect genuine outcry from this unless it gets made into a hyperbolic TikTok.

New comment by bigyabai in "Qwen3.6-35B-A3B: Agentic coding power, now open to all"

bigyabai — Thu, 16 Apr 2026 17:58:38 +0000

Your current laptop is still a fine thin client. Unless you program in the woods, it's probably cheapest to build a home inference box and route it over Tailscale or something.

New comment by bigyabai in "Qwen3.6-35B-A3B: Agentic coding power, now open to all"

bigyabai — Thu, 16 Apr 2026 15:56:33 +0000

LPDDR is LPDDR. There's nothing "unified" about it architecturally.

New comment by bigyabai in "Ask HN: Suggest good RSS Readers please?"

bigyabai — Thu, 16 Apr 2026 15:52:47 +0000

I'm fond of NewsFlash on Linux: https://flathub.org/en/apps/io.gitlab.news_flash.NewsFlash

New comment by bigyabai in "Qwen3.6-35B-A3B: Agentic coding power, now open to all"

bigyabai — Thu, 16 Apr 2026 15:38:18 +0000

taps the sign

  Unified Memory Is A Marketing Gimmeck. Industrial-Scale Inference Servers Do Not Use It.

New comment by bigyabai in "Qwen3.6-35B-A3B: Agentic coding power, now open to all"

bigyabai — Thu, 16 Apr 2026 15:32:58 +0000

Seconding this. You can get A3B/A4B models to run with 10+ tok/sec on a modern 6/8GB GPU with 32k context if you optimize things well. The cheapest way to run this model at larger contexts is probably a 12gb RTX 3060.

New comment by bigyabai in "Solar panels are forming rainfall clouds and oases in the middle of deserts"

bigyabai — Wed, 15 Apr 2026 22:21:39 +0000

Blogspam version of this article: https://www.science.org/content/article/massive-solar-farms-...

New comment by bigyabai in "CPUs Aren't Dead. Gemma2B Out Scored GPT-3.5 Turbo on Test That Made It Famous"

bigyabai — Wed, 15 Apr 2026 18:02:03 +0000

But GPT-3.5 was benchmaxxing too.

New comment by bigyabai in "Fuck the cloud (2009)"

bigyabai — Wed, 15 Apr 2026 18:00:28 +0000

You'd be surprised. There are tens of thousands (probably millions?) of people that happily use NFS or hosted services on a bog-standard Windows device with no gripes.

New comment by bigyabai in "New search engine reveals if ancestors were in Nazi party"

bigyabai — Wed, 15 Apr 2026 17:49:12 +0000

"Pistol duel at dawn - de_nuke."

New comment by bigyabai in "Google Gemma 4 Runs Natively on iPhone with Full Offline AI Inference"

bigyabai — Wed, 15 Apr 2026 17:36:50 +0000

The ANE is not a fast or realistic way to infer modern LLMs.

New comment by bigyabai in "Google Gemma 4 Runs Natively on iPhone with Full Offline AI Inference"

bigyabai — Wed, 15 Apr 2026 17:34:24 +0000

> No amount of bloat matches what an LLM needs.

I don't think that's necessarily true. For instance, LinkedIn uses more memory than Gemma E2B inference does.

New comment by bigyabai in "Google Gemma 4 Runs Natively on iPhone with Full Offline AI Inference"

bigyabai — Wed, 15 Apr 2026 14:12:40 +0000

Nothing that Mac and Windows users aren't already used to.

New comment by bigyabai in "Keep Android Open"

bigyabai — Wed, 15 Apr 2026 13:42:56 +0000

We have other mobile OSes, even ones that support Android apps like Jolla and PostmarketOS. People don't use them.