Hacker News: kroaton

New comment by kroaton in "Developer gets Half-Life running at 30 FPS on a Nokia N95"

kroaton — Mon, 15 Jun 2026 17:42:04 +0000

This has been answered many times already. The short version: Brave's adblocker is not an extension, so manifest v3 has no effect on it.

New comment by kroaton in "Anthropic apologizes for invisible Claude Fable guardrails"

kroaton — Thu, 11 Jun 2026 19:52:40 +0000

To be fair, GPT5.5-Xhigh is similarly capable and has not burned the world down.

New comment by kroaton in "Anthropic surpasses OpenAI to become most valuable AI startup"

kroaton — Sat, 30 May 2026 22:50:40 +0000

Anthropic nuked a big chunk of that "developer sentiment" when they rug-pulled us with the rate limits and gaslit us with "it was just a bug, guys!".

New comment by kroaton in "Reddit blocks JSON API access, which leaves read-only apps like RDX unusable"

kroaton — Sat, 30 May 2026 14:59:18 +0000

How is this not getting any traction? This is a massive problem. Pretty much every Reddit news app out there is now dead.

New comment by kroaton in "Qwen3.7-Max: The Agent Frontier"

kroaton — Wed, 20 May 2026 21:15:43 +0000

Did you even use it? It was nerfed to hell and back. It stopped following instructions, forgot what sub-agents responded and so on. Stop spreading this pro-Anthropic narrative. They did a rug pull due to lack of compute.

New comment by kroaton in "Zero-native – Build native desktop apps with web UI"

kroaton — Wed, 13 May 2026 11:25:57 +0000

If Tauri ever gets proper webgpu support, that'll be the Electron killer.

New comment by kroaton in "DeepSeek V4 – almost on the frontier"

kroaton — Sat, 02 May 2026 12:49:31 +0000

Claude Code poisons non-anthropic models in usage. We found this out when the code was leaked. Use a fork or OpenCode/pi-coding-agent

New comment by kroaton in "DeepSeek v4"

kroaton — Fri, 24 Apr 2026 12:15:38 +0000

Ask western models about Israel's genocides and mass rapes in Palestine, Lebanon, etc.

New comment by kroaton in "Qwen3.6-27B: Flagship-Level Coding in a 27B Dense Model"

kroaton — Wed, 22 Apr 2026 20:38:10 +0000

Buy any Strix Halo box and have fun with your 128GB of VRAM.

New comment by kroaton in "Qwen3.6-27B: Flagship-Level Coding in a 27B Dense Model"

kroaton — Wed, 22 Apr 2026 16:09:32 +0000

A3B-35B is better suited for laptops with enough VRAM/RAM. This dense model however will be bandwidth limited on most cards.

The 5090RTX mobile sits at 896GB/s, as opposed to the 1.8TB/s of the 5090 desktop and most mobile chips have way smaller bandwith than that, so speeds won't be incredible across the board like with Desktop computers.

New comment by kroaton in "$500 GPU outperforms Claude Sonnet on coding benchmarks"

kroaton — Fri, 27 Mar 2026 13:26:46 +0000

For autocomplete, Qwen 3.5 9B should be enough even at Q4_k_m. The upcoming coding/math Omnicoder-2 finetune might be useful (should be released in a few days).

Either that or just load up Qwen3.5-35B-A3B-Q4_K_S I'm serving it at about 40-50t/s on a 4070RTX Super 12GB + 64GB of RAM. The weights are 20.7GB + KV Cache (which should be lowered soon with the upcoming addition of TurboQuant).

New comment by kroaton in "Moving from GitHub to Codeberg, for lazy people"

kroaton — Thu, 26 Mar 2026 21:40:10 +0000

I did the same a few months ago when I read that multiple big OSS Linux projects were moving to it and it's been phenomenal so far.

New comment by kroaton in "Flash-MoE: Running a 397B Parameter Model on a Laptop"

kroaton — Sun, 22 Mar 2026 22:23:16 +0000

It could just as easily be a $3000-4000 Strix Halo laptop.

New comment by kroaton in "Big data on the cheapest MacBook"

kroaton — Thu, 12 Mar 2026 21:45:37 +0000

If SPTM is active on the chip, we are not going to be getting Linux at all.

New comment by kroaton in "Gaussian Splatting – A$AP Rocky "Helicopter" music video"

kroaton — Sun, 18 Jan 2026 21:13:11 +0000

Loving this; great work! Do you talk about the process anywhere in more depth?

New comment by kroaton in "Ask HN: What did you find out or explore today?"

kroaton — Thu, 15 Jan 2026 14:00:54 +0000

I remember thinking the same thing and this articles goes over most of the arguments here - https://milvus.io/blog/why-im-against-claude-codes-grep-only...

When it came out and I think it was Boris from Anthropic that said they experimented a lot with Vector Search and grep just worked better.

You can try it out using the Claude-Context MCP - https://github.com/zilliztech/claude-context

New comment by kroaton in "Anthropic Explicitly Blocking OpenCode"

kroaton — Thu, 15 Jan 2026 02:19:07 +0000

I hope the upcoming DeepSeek coding model puts a dent in Anthropic’s armor. Claude 4.5 is by far the best/fastest coding model, but the company is just too slimy and burning enough $$$ to guarantee enshitification in the near future.

New comment by kroaton in "Nvidia to buy assets from Groq for $20B cash"

kroaton — Thu, 25 Dec 2025 11:31:26 +0000

Cerebras currently has GLM4.6 on it, and will be getting GLM4.7 soon.

New comment by kroaton in "Google's year in review: areas with research breakthroughs in 2025"

kroaton — Wed, 24 Dec 2025 22:01:15 +0000

Lol.

New comment by kroaton in "GLM-4.7: Advancing the Coding Capability"

kroaton — Tue, 23 Dec 2025 13:09:38 +0000

That's gone now. They do drops from time to time, but their compute platform is saturated.