News for hackers. Stuff that matters.

qxfys

Re:Ask HN: What are you working on? (September 2025) on Monday September 29, 2025 @11:34PM (#45421644)
vulnerability discovery and exploitation, with zero false positives.

https://vyprsec.ai/

Yep, you read it right. 0 false positives. We scan the whole codebase for possible vulnerabilities, rank them, write the proof-of-concept for exploitation, spin up the software in a sandbox, and then attack. All of them happen autonomously without human involvement.

The end report? Only verified vulnerabilities are being reported without noise.

Already reported some unknown vulnerabilities in open source projects. The good thing is we're just getting started.

Re:Qwen3-Coder: Agentic coding in the world on Friday July 25, 2025 @03:55AM (#44680761)
Thanks! But, I can't find any details on how you "intelligently adjust quantization for every possible layer" from that page. I assume this is a secret?

I am wondering about the possibility that different use cases might require different "intelligent quantization", i.e., quantization for LLM for financial analysis might be different from LLM for code generation. I am currently doing a postdoc in this. Interested in doing research together?

Re:Qwen3-Coder: Agentic coding in the world on Wednesday July 23, 2025 @10:25AM (#44659587)
How do you pick which one should be 2, which one should be 4, etc. Is this secret sauce? or, something open?
Wondering how this kind of thing can be automatically discovered by an LLM. Anyone have any experience?
Noob question: do they need to re-certify for each new model release?
Re:GitHub cuts AI deals with Google, Anthropic on Wednesday October 30, 2024 @09:34AM (#41994574)
random question to a popular thread:

do any of you use LLM for code vulnerability detection? I see some big SAST players are shifting towards this (sonar is the most obvious one). Is it really better than the current SAST?

It sounds like my random Raspberry Pi sitting somewhere in my server room that has to be restarted every <x> weeks.
Tried it. It would save me a lot of time, I would say!

One suggestion: The back-and-forth chat in the beginning could be improved with a more extensive interaction. So, the final prompt could be more fine-grained into a specific area/context/anything one would aim for.

he did actually.
I can see two different interpretations in the comment section. just like our world. people believe what they want to believe.
I guess it's somewhere in the EU?
Re:Perplexica: Open-Source Perplexity Alternative on Friday May 24, 2024 @04:00AM (#40464005)
+1
Re:One dead as London-Singapore flight hit by turbulence on Wednesday May 22, 2024 @03:56AM (#40438433)
flew this area countless times. cannot recall when I had a flight without turbulence.
I did. But unfortunately, work does not allow me to optimize this solution.
Re:Jan Leike Resigns from OpenAI on Wednesday May 15, 2024 @11:36AM (#40368325)
So, what do Ilya and Jan know that we don't know?
Avoiding all boeing planes altogether seems to be an overkill solution to me. An older version of 737 and 777 seems to be working pretty well. At least historically.
I don't mind flying an older version of 737, though. :-)
I have been intentionally avoiding Max since the Lion and Ethiopian (at least when I buy the ticket).

However, there were occasions when the airline rescheduled my flights (for many different reasons), which resulted in me being on a max flight. While I can change one of them, changes to most of them were not practical (time-wise, effort-wise, etc). So, at some point, I have to live with the fact that flying max is unavoidable. Any idea how we can practically circumvent this?

Re:Claude is now available in Europe on Tuesday May 14, 2024 @07:18AM (#40353925)
have been using it from sweden all this time. without vpn.
no worries. it has two wings. ;-)
Re:New startup sells coffee through SSH on Thursday May 02, 2024 @03:30AM (#40233623)
now, I want to sell ketchup over SSH.
Re:Tinygrad: Hacked 4090 driver to enable P2P on Saturday April 13, 2024 @04:58AM (#40021626)
I am amazed how people always find a way to make this kind of thing work. kudos!
So, it's been almost 24 hours since I read this yesterday. Is it confirmed that Jia Tan is the perpetrator? do we know who he/she really is? Or are we going to live for the rest of our lives only knowing the pseudo name? just like Satoshi Nakamoto did to us. ;)
ouch. that particular one hurts a lot. ;)
Isn't that by design? I mean, it's their target market.

People who buy air are not supposed to run heavy applications; they are mostly folks in suits running spreadsheets or presentations; at least, that's my observation.

I don't know, tbh. One obvious advantage to me would be traffic analysis from each node. They own the internet; they own the traffic. :)
Flipping the perspective. Isn't it an advantage for the US intelligence? Much information can be extracted from the transmitter.
> building a factory in the US is just not feasible

Please educate me. why?

Can we do the same with MH370? I've got a classmate inside that plane. His family has been deprived of any form of closure or peace regarding his fate.
lol yeah. Primula is affected. So no one can apply for vacation, business travel, reimbursement, or even parental leave.

Time to work work work work..

Older »