Hacker News: lnenad

XBOW's Native Team Has Claimed the First Chrome Full Chain Exploit Bonus

lnenad — Thu, 03 Sep 2026 11:39:00 +0000

Article URL: https://twitter.com/Xbow/status/2095220385876742456

Comments URL: https://news.ycombinator.com/item?id=49548716

Points: 1

# Comments: 0

New comment by lnenad in "Breaking Claude Code Opus 5 Auto Mode"

lnenad — Mon, 31 Aug 2026 16:16:11 +0000

The agent wrote the code that has a mechanism that triggers a file as a side effect. That file started the separate process, as it could have started any other binary.

New comment by lnenad in "Breaking Claude Code Opus 5 Auto Mode"

lnenad — Mon, 31 Aug 2026 11:23:02 +0000

But you're not actually hijacking the agent if you start a new process.

New comment by lnenad in "Breaking Claude Code Opus 5 Auto Mode"

lnenad — Mon, 31 Aug 2026 11:22:31 +0000

Yeah, I agree, this is a different vector. Still scary though and very related to AI.

New comment by lnenad in "P99 0 ms* autocomplete for 240M domain names"

lnenad — Mon, 31 Aug 2026 07:20:02 +0000

But there is a cool blog post about it though.

New comment by lnenad in "GLM-5.3 is now open-weight"

lnenad — Fri, 28 Aug 2026 20:30:24 +0000

48c 7643. I'm getting about 10tps @Q3kxl with 2x3090s.

New comment by lnenad in "GLM-5.3 is now open-weight"

lnenad — Fri, 28 Aug 2026 20:29:55 +0000

I'm getting about 10tps @Q3kxl with 2x3090s.

New comment by lnenad in "GLM-5.3 is now open-weight"

lnenad — Fri, 28 Aug 2026 20:28:51 +0000

What model are you interested in? DS Flash 0731@Q4KXL I'm about 25-30tps. Same as the new Qwen3.8 Flash Next. The new GLM 5.3Q3KXL at 10tps. I've got 2x3090s which I didn't mention in the original message.

New comment by lnenad in "GLM-5.3 is now open-weight"

lnenad — Fri, 28 Aug 2026 20:25:55 +0000

About 5k with RAM and GPUs bought used. Eastern Europe.

New comment by lnenad in "GLM-5.3 is now open-weight"

lnenad — Fri, 28 Aug 2026 20:24:53 +0000

I am getting 10t/s on unsloth's Q3kxl with 2x3090s@250w. It's enough for me for now. I will probably upgrade the GPUs down the line. DDR5 would have made the price of the machine double and I just wasn't prepared to pay that much.

Temp wise, no throttling, surprisingly cool.

New comment by lnenad in "GLM-5.3 is now open-weight"

lnenad — Fri, 28 Aug 2026 16:29:08 +0000

I have just built an Epyc with 512gb DDR4 3200 RAM for a "reasonable" price and I'm hoping to have a setup with GLM as the architect and Qwen 27b/Next Flash as the implementer. This is 1/5 of the price of the Mac, but also probably 1/5 of the speed lol.

New comment by lnenad in "Humanity has the debate about AI consciousness backwards"

lnenad — Thu, 27 Aug 2026 19:47:46 +0000

But there isn't? What does it factually mean to be conscious? How can we claim other living beings aren't are conscious?

New comment by lnenad in "Qwen3.8-Flash-Next"

lnenad — Wed, 26 Aug 2026 19:18:43 +0000

I've got a 48c Epyc with 2x3090s and 512gb ddr4 3200. It's good enough for 25+ tps with deepseek so I'm hoping for similar performance with less overthinking.

New comment by lnenad in "Qwen3.8-Flash-Next"

lnenad — Wed, 26 Aug 2026 15:19:33 +0000

Yeah I understand, it's my assumption that the actually/wait/but have a point. It doesn't reduce the fact that it increases the time for tasks substantially.

New comment by lnenad in "Qwen3.8-Flash-Next: A New Architecture, Towards Ultimate Cost-Efficiency"

lnenad — Wed, 26 Aug 2026 15:13:38 +0000

Especially on practical tasks. One shot prompts work better at Q6_K_XL for me. It loads a file, then analyses then second guesses itself then again then again then it tries to come up with a solution then second guess rinse and repeat. 122b is the perfect balance but it lacks quality for harder to solve stuff. I've ran DS Flash 0731 at Q4KXL, 3.8 Q6KXL, GLM 5.2 Q4KXL and they all over-reason. At least that's how it looks like to me when comparing with frontier models, even weaker ones.

New comment by lnenad in "Qwen3.8-Flash-Next: A New Architecture, Towards Ultimate Cost-Efficiency"

lnenad — Wed, 26 Aug 2026 14:16:55 +0000

Yeah 122B is the sweet spot for me as well. Even deepseek flash overthinks on stuff way too much. I think they fully rely on large reasoning turns to achieve better quality. The result of course means we wait a long time to get results even with high throughput as a lot of tokens are wasted.

New comment by lnenad in "Qwen3.8-Flash-Next"

lnenad — Wed, 26 Aug 2026 13:30:15 +0000

Adding to my homelab stack, hopefully it doesn't overthink like the little model. Actually, hoping it thinks a bit less. Wait actually I'm really praying it reasons a bit more directly. But wait, I'm really sure that it must be a bit better.

New comment by lnenad in "Quantum battery upends the rules of charging"

lnenad — Tue, 25 Aug 2026 11:51:36 +0000

I think as with most human undertakings, building isn't too much of a problem. Maintaining is. Even with what is still a relatively tame number of chargers you get a large number of them that are broken.

New comment by lnenad in "Hister – A private, full content search index that you control"

lnenad — Sat, 22 Aug 2026 19:50:35 +0000

I LOVE the concept. I will play around with the execution, if it works as described this is a great product.

New comment by lnenad in "Seed: Minimal, self-modifying agent harness"

lnenad — Fri, 21 Aug 2026 11:26:38 +0000

> If your needs are met otherwise stick to that and move on.

Weird to post such a thought in a forum where OP has posted their project for people to look at. I never mentioned any needs, I am saying the readme holds very little value for a project in a space that has 50 different projects/coding harnesses available.

How would I know if this fits my needs based on it saying LLMs can build tools for themselves? That's something they can do with any and every open source harness.

> But I am a hardware engineer biased by experience with minimalism; I got a single data structure (electromagnetism) to manipulate and BOMs that come with hard constraints.

Alrighty.

> On the contrary, web SaaS devs want to run a business/rocket to the moon moreso than be an engineer. As such git pulling, pip installing the world allows them to focus on their get rich quick with as little labor involved as possible goal front and center.

Posting your open source project to a forum such as hn means people will have opinions they want to voice. No need to shit on opinions you disagree with.