Hacker News: StrauXX

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

StrauXX — Wed, 27 May 2026 17:59:42 +0000

Algorithms are also improving. I believe it's very unlikely for these two improvements together to not result in one to two orders of magnitude cheaper cost per "intelligence". Of course, that might just make use cases that are too expensive today viable and thereby increase usage further.

New comment by StrauXX in "Frontier AI has broken the open CTF format"

StrauXX — Sat, 16 May 2026 10:21:01 +0000

LLMs don't tend to help much when solving challenges beyond their skill level. Either they one-shot a challenge, or thei are almost useless as a companion for them.

New comment by StrauXX in "Frontier AI has broken the open CTF format"

StrauXX — Sat, 16 May 2026 10:17:33 +0000

It is a hard requirement. Once you reach higher levels of challenges you spend most of your time reading through RFCs, web sepcs, Github issues, mailing lists, papers, random bugtrackers and library/framework code. There is no way to create a whitelist for that. Besides, a firewall won't stop good hackers.

New comment by StrauXX in "A recent experience with ChatGPT 5.5 Pro"

StrauXX — Sat, 09 May 2026 09:21:23 +0000

Which indications are that?

New comment by StrauXX in "AI should elevate your thinking, not replace it"

StrauXX — Mon, 27 Apr 2026 13:04:51 +0000

It is definetly cheaper now. What I want to say with this, is that token costs rising so dramatically that AI usage becomes uneconomical is not a high probability future. Even if AI subscriptions were sold heavily below cost (which is also unlikely, after R&D).

New comment by StrauXX in "AI should elevate your thinking, not replace it"

StrauXX — Sun, 26 Apr 2026 21:11:56 +0000

Self hosting at a reasonable scale is much cheaper than people think. I am running clusters of DGX Spark machines with BiFrost load balancers in our company and for client projects. They work flawlessly!

128 GB unified memory, Nvidia chip and ARM CPU for just around 3k€ net. They easily push ~400 input and ~100 output tokens per second per device on say gpt-oss-120b. With two devices in a cluster, thats enough performance for >20 concurrent RAG users or >3 "AI augmented" developers.

And they don't even pull that much power.

New comment by StrauXX in "The Tech Oligarch's Republic"

StrauXX — Wed, 22 Apr 2026 17:18:26 +0000

The party is called the "Christian Democratic Party" but in practice pushes no christian policies. 47% of germans are legally atheists. Only 5% regularly visit mass.

New comment by StrauXX in "NSA is using Anthropic's Mythos despite blacklist"

StrauXX — Mon, 20 Apr 2026 13:43:40 +0000

He has changed his opinion completely. Yes, the ratio has turned.

New comment by StrauXX in "Stripe's Payment APIs: the first 10 years (2020)"

StrauXX — Mon, 20 Apr 2026 10:30:06 +0000

Only Stripe offers a service doing international VAT for you.

New comment by StrauXX in "The local LLM ecosystem doesn’t need Ollama"

StrauXX — Thu, 16 Apr 2026 08:12:50 +0000

vLLM isn't suitable for people running LLMs side-by-side with regular applications on their PC. It is very good at hosting LLMs for production on dedicated servers. For the prod usecase ollama/llamacpp are practically useless (but that's ok - it's not the projects goal to be).

New comment by StrauXX in "Show HN: Every CEO and CFO change at US public companies, live from SEC"

StrauXX — Wed, 15 Apr 2026 15:13:37 +0000

Northdata [1] does basically this, but mostly ingests European data currently. Perhaps they will expand to ingest US data as well at some point. Not affiliated with them in any way. I just use them to look into company structures every now and again.

[1] https://www.northdata.com/

New comment by StrauXX in "N-Day-Bench – Can LLMs find real vulnerabilities in real codebases?"

StrauXX — Tue, 14 Apr 2026 05:08:06 +0000

Do you plan on adding more models in the future? I would love to see how other OSS modles like Gemma, GPT-OSS and Qwen fare.

New comment by StrauXX in "Small models also found the vulnerabilities that Mythos found"

StrauXX — Sat, 11 Apr 2026 23:57:26 +0000

A lot of comments here are dismissing this post because the relevant code was isolated. But thats the exact same thing Anthropic did with Mythos! They describe their (very lean) harness in the Anthropic Red Mythos blog post. The harness first assigns each file in the given codebase an importance value. Then points claude code at the cpdebase with a prompt stating that it should focus on that file. It spawns a claude code instances for each file in the codebase.

So no, the fact that the posters isolated the relevant code does not invalidate their findings.

[1] https://red.anthropic.com/2026/mythos-preview/

New comment by StrauXX in "Project Glasswing: Securing critical software for the AI era"

StrauXX — Wed, 08 Apr 2026 15:20:02 +0000

There are quite a few such startups already out there. Results are mixed so far. Though I believe they get much better over the coming months and years.

New comment by StrauXX in "Why Switzerland has 25 Gbit internet and America doesn't"

StrauXX — Mon, 06 Apr 2026 08:06:11 +0000

Switzerland is one of the least homogenous countries in Europe! Four languages, relatively weak federal government with strong local (canton) governments. Most state services are handled on a localized (canton) level.

New comment by StrauXX in "LibreOffice – Let's put an end to the speculation"

StrauXX — Sun, 05 Apr 2026 21:11:33 +0000

I had the same impression. The introduction reads really unprofessionally.

AI Wafer Timeline

StrauXX — Tue, 31 Mar 2026 22:43:57 +0000

Article URL: https://xcancel.com/aakashgupta/status/2038813799856374135

Comments URL: https://news.ycombinator.com/item?id=47594448

Points: 2

# Comments: 0

Organizational Politics and the Security Program

StrauXX — Sun, 29 Mar 2026 19:16:08 +0000

Article URL: https://www.philvenables.com/post/organizational-politics-the-security-program

Comments URL: https://news.ycombinator.com/item?id=47566221

Points: 3

# Comments: 0

New comment by StrauXX in "My minute-by-minute response to the LiteLLM malware attack"

StrauXX — Sat, 28 Mar 2026 08:48:20 +0000

If your requirements are just to load balance between selhosted AI servers: nginx. If you want a more thorough system with configurability, logging, etc.: Bifrost from MaximAI

New comment by StrauXX in "My minute-by-minute response to the LiteLLM malware attack"

StrauXX — Fri, 27 Mar 2026 08:28:42 +0000

LiteLLM isn't a good choice for a proxy in any case. It introduces a lot of lag amd latency and the features are often half baked. To me, it looks like a vibecoded application without a product owner. And the code itself isn't very organized either. I evaluated it for a project a few months ago and will never use it for anything production. Theres a few much better alternatives out there.