Hacker News: 3abiton

New comment by 3abiton in "Local Qwen isn't a worse Opus, it's a different tool"

3abiton — Thu, 18 Jun 2026 07:57:10 +0000

And a big thing that's missing is ... the harness comparison. Ot plays a very big role. I use forge, and I have been inpressed with what it can do given all the limitations of local models.

New comment by 3abiton in "Ask HN: Has anyone replaced Claude/GPT with a local model for daily coding?"

3abiton — Mon, 15 Jun 2026 23:41:48 +0000

I think nearly everyone mentioned Qwen, so my turn I guess. Qwen 3.6 35B Q8 (MTP), on a Strix Halo, with llama.cpp. Around 40-50 t/s. Really great pefromance, I get always suprised by its capability. I used with forge-code directly in zsh. For long context 150k+) it start degrading and forgetting.

New comment by 3abiton in "A backdoor in a LinkedIn job offer"

3abiton — Mon, 15 Jun 2026 23:33:20 +0000

I have the same. The difference is, if you do email verification, you will "verified" status. If not, you can still add the company to your linkedin, just unverified, which is not a label.

New comment by 3abiton in "Gemma 4 QAT models: Optimizing compression for mobile and laptop efficiency"

3abiton — Fri, 05 Jun 2026 23:48:28 +0000

Are there evidence that this approach helps maintain "accuracy" performance when quantized? It sounds a bit like mxfp4 with gpt-oss, which was a confusing model upon release.

New comment by 3abiton in "Codex just found a "workaround" of not having sudo on my PC"

3abiton — Sun, 31 May 2026 22:28:01 +0000

Podman has lots of underappreciated features, and it's fully open-source!

New comment by 3abiton in "I think Anthropic and OpenAI have found product-market fit"

3abiton — Thu, 28 May 2026 05:29:24 +0000

Not to mention the competition: chinese open-weight models and open-source harnesses. Qwen3.6-(27B and 35B) have proven to be worthy and capable of running locally. I am confident more SMEs would look into this as a solution given the ballooning costs of API usage. You get a decent setup with an RTX 6000 Pro.

New comment by 3abiton in "An OpenAI model has disproved a central conjecture in discrete geometry"

3abiton — Wed, 20 May 2026 22:25:25 +0000

> * That can still yield useful "discoveries" in certain fields, absent the discovery of new mechanics that exist outside said training data

One can argue, new knowledge is just restructured data.

I think the main concerns about LLMs is the inherent "generative" aspects leading to hallucinations as a biproduct, because that's what produces the noi. Joint Embedding approaches are rather an interesting alternative that try to overcome this, but that's still in research phase.

Reverse engineering Android malware from popular Chinese projectors

3abiton — Mon, 18 May 2026 00:36:10 +0000

Article URL: https://zanestjohn.com/blog/reing-with-claude-code

Comments URL: https://news.ycombinator.com/item?id=48174465

Points: 89

# Comments: 19

New comment by 3abiton in "I hate soldering"

3abiton — Mon, 11 May 2026 21:54:14 +0000

Step 1: Have a workshop space Step 2: ? Step 3: Profit

New comment by 3abiton in "Granite 4.1: IBM's 8B Model Matching 32B MoE"

3abiton — Thu, 30 Apr 2026 16:36:43 +0000

> Qwen3.6 35b a3b is still my local champion but I may use this for auto complete and small tasks.

I second this! Using the Unsloth Q6 (I forgot the exact name). Currently using it with forgecode (with zsh), on my Strix Halo, and it's suprisingly really good. I would say slightly Similar to Haiku 4.5, plus additional privacy, minus speed. It's surprisingly really fast for the hardware, given the speculative decoding, still PP is on the slow side.

New comment by 3abiton in "Tell HN: Fiverr left customer files public and searchable"

3abiton — Thu, 16 Apr 2026 07:50:24 +0000

They patched the "non-existent" issue it seems. And totally denied it happened in the first place. Honestly, someone should do a dump of redacted client documents to teach them a lesson. Short of a class action lawsuit would be an understatement. This is really huge.

New comment by 3abiton in "The Orange Pi 6 Plus"

3abiton — Tue, 14 Apr 2026 22:02:29 +0000

Even though all can be replaced by a decent mini pc with beefy memory, with lots of VMs.

New comment by 3abiton in "Seven countries now generate nearly all their electricity from renewables (2024)"

3abiton — Sun, 12 Apr 2026 21:17:18 +0000

> So, the lessons for all other countries in the world is pretty clear: grow yourselves some mountains, dig yourselves a big river, and dam, baby, dam !!

You're forgetting corruption. Many countries can easily go 100% renewable, but there is no profit for dictators/politicians to do so. Most of africa, or the middle east, yet you still have many regions without electricity or water, so that people worry about food for tomorrow instead of better governance in the future.

New comment by 3abiton in "How We Broke Top AI Agent Benchmarks: And What Comes Next"

3abiton — Sun, 12 Apr 2026 12:31:29 +0000

Benchmarking has been already known to be far from a signal of quality for LLMs, but it's the "best" standardized way so far. Few exists like the food truck and the svg test. At the end of the day, there is only 1 way: having your own benchmark for your own application.

New comment by 3abiton in "Installing every* Firefox extension"

3abiton — Sat, 11 Apr 2026 07:23:04 +0000

> Dr. B is the king of slop, with 84 extensions published, all of them vibe coded. > How do I know? Most of their extensions has a README.md in them describing their process of getting these through addon review, and mention Grok 3. Also, not a single one of them have icons or screenshots. > Personally, I’m shocked this number is this low. I expected to see some developers with hundreds!

This is really surprising. Either because Firefox is not that popular ir mozilla has an automatic filter?

New comment by 3abiton in "What is RISC-V and why it matters to Canonical"

3abiton — Fri, 10 Apr 2026 20:25:22 +0000

While its current performance is not competitive, there are currently interesting options. I got the orange pi riscv version, mainly to test riscv while it's slow compared to other arm socs, it's still better than I expected. There are even risc v TPUs now.

New comment by 3abiton in "Research-Driven Agents: When an agent reads before it codes"

3abiton — Fri, 10 Apr 2026 15:55:30 +0000

I am surprised you found RST better than markdown.

New comment by 3abiton in "Open Source Security at Astral"

3abiton — Thu, 09 Apr 2026 06:55:17 +0000

I don't see the connection though?

New comment by 3abiton in "Show HN: Real-time AI (audio/video in, voice out) on an M3 Pro with Gemma E2B"

3abiton — Wed, 08 Apr 2026 15:32:41 +0000

We're getting closer now to black mirror level of technology.

New comment by 3abiton in "$500 GPU outperforms Claude Sonnet on coding benchmarks"

3abiton — Wed, 08 Apr 2026 15:23:08 +0000

The method here is model agnostic.