Hacker News: catlifeonmars

New comment by catlifeonmars in "Solo – a .so loader for static Linux binaries"

catlifeonmars — Wed, 19 Aug 2026 00:39:39 +0000

So not completely static, since it must link against a libc :P

New comment by catlifeonmars in "How to disable or avoid intrusive AI"

catlifeonmars — Tue, 18 Aug 2026 02:59:24 +0000

It’s shorthand for “makes you cringe” :shrug:

New comment by catlifeonmars in "Software Engineering fundamentals matter more"

catlifeonmars — Sun, 16 Aug 2026 19:43:13 +0000

> The only situation where I didn't see this happening was in startups where people are actually busy.

FWIW you still see this, it just tends to not survive for too long.

New comment by catlifeonmars in "What happens when an LLM never sees material beyond fifth grade?"

catlifeonmars — Sun, 16 Aug 2026 19:18:09 +0000

Does it not know how to say “I don’t know”?

New comment by catlifeonmars in "Does anyone run Postgres without PgBouncer?"

catlifeonmars — Sun, 16 Aug 2026 18:59:23 +0000

That’s ok if the number of workers is bounded and small. But if you’re operating on the order of hundreds to thousands of workers, maybe not so much.

New comment by catlifeonmars in "Show HN: I built a native app for coding agents with Rust and GPUI"

catlifeonmars — Sun, 16 Aug 2026 04:16:48 +0000

The lack of commas in the title yields some ambiguity:

- is it an app for creating custom agents?

- do the agents use Rust and GPUI?

New comment by catlifeonmars in "A controversial Alzheimer's surgery is said to reverse symptoms"

catlifeonmars — Sat, 15 Aug 2026 18:20:32 +0000

Hanlon’s razor much?

New comment by catlifeonmars in "Google is making private AI practical with homomorphic encryption"

catlifeonmars — Sat, 15 Aug 2026 04:29:09 +0000

They’re not equivalent. One provides a best effort guarantee and is not trivial to implement correctly (it’s the Secure Enclave). I agree that you _could_ do all those things and do them correctly. However there are a lot more points of failure.

E2E encryption (including homomorphic encryption) have the nice property that there are much fewer ways for things to fail.

(Tangentially, attestation is basically trying to ensure that faults are obvious, but that doesn’t reduce the probability of the faults in the first place).

New comment by catlifeonmars in "Google is making private AI practical with homomorphic encryption"

catlifeonmars — Sat, 15 Aug 2026 04:00:29 +0000

> aren’t there more efficient ways

Heh yes absolutely, but there is some nuance.

Secure Enclave still requires you to trust the operator and also trust that it’s configured properly, supply chain is secure, etc.

The beauty of FHE is that it doesn’t rely on the compute being secure. All you need to secure are things you already have control over as a client.

I agree with you it’s still way too slow to be generally useful. (By general, I mean practical for arbitrary computation — you can relax the requirement and have fast homomorphic encryption if you only do specific kinds of operations).

New comment by catlifeonmars in "Google is making private AI practical with homomorphic encryption"

catlifeonmars — Sat, 15 Aug 2026 03:52:34 +0000

If the bottleneck is data transfer rate or volume, then I think you’ll find more use from today’s FHE. Incidentally there is a fair amount of overlap in the use case space between homomorphisms and oblivious transfer.

New comment by catlifeonmars in "How I use LLMs to learn complex topics"

catlifeonmars — Sun, 09 Aug 2026 20:16:24 +0000

I don’t agree with the implication that there has to be a practical reason to learn new things. I enjoy learning new technologies because it’s fun.

New comment by catlifeonmars in "LLMs won't break symmetric crypto"

catlifeonmars — Thu, 06 Aug 2026 03:03:46 +0000

Agreed that many of the articles claims are a bit weak. One point is reasonably strong though: symmetric crypto may not be breakable (battle tested).

New comment by catlifeonmars in "LLMs won't break symmetric crypto"

catlifeonmars — Thu, 06 Aug 2026 02:54:48 +0000

Perhaps, but it’s still trivially easy to increase the difficulty of factorization problems on classical computers, We need a machine that can run Shor’s algorithm before integer factorization is practical and we’re still a long way out f M that.

New comment by catlifeonmars in "Octane – React’s programming model, compiled"

catlifeonmars — Mon, 03 Aug 2026 16:07:57 +0000

Code quality at one point in time is useful, but what you want is dependability over time and responsiveness to security fixes; this means looking at the maintainer and their attention to detail.

Sure, the website by itself doesn’t mean anything, but placed into context it could mean the maintainer doesn’t put much effort into clear, unambiguous communication.

New comment by catlifeonmars in "Prevent cognitive debt by manually retyping LLM-generated code"

catlifeonmars — Mon, 03 Aug 2026 15:18:24 +0000

Don’t good generals know how everyone else in their army does their jobs? At the very least to the extent they can see through the bullshit.

For example, I don’t handwrite the code that the compiler produces, but I know how to troubleshoot the IR/assembly and fix optimization issues that arise occasionally. That makes me better at my job of directing the compiler.

New comment by catlifeonmars in "When random.bytes() runs but doesn't work"

catlifeonmars — Sun, 02 Aug 2026 15:22:26 +0000

A lot of the vulnerabilities being discovered are low hanging fruit. An LLM could discover them, but so could a human.

Let me amend your statement:

If the industry most seasoned experts are not engaging in regular security reviews of their firmware, they are not experts, but clowns.

LLMs are useful for this but not absolutely necessary to address the core issue.

New comment by catlifeonmars in "GitHub is the wrong shape for this new world"

catlifeonmars — Thu, 30 Jul 2026 01:39:38 +0000

> I'm not going to pretend I know how to do Github better than Github

To be fair, there are lots of forges that are better than GitHub; knowing how to GitHub better than GitHub is an extremely low bar.

New comment by catlifeonmars in "Document-borne AI worms can self-propagate through Copilot for Word"

catlifeonmars — Wed, 29 Jul 2026 13:14:44 +0000

People who use machines based on the von Neumann architecture?

New comment by catlifeonmars in "About the security content of macOS Tahoe 26.6"

catlifeonmars — Tue, 28 Jul 2026 14:53:08 +0000

How would you enforce a single implementation of path parsing?

New comment by catlifeonmars in "Our position on open-weights models"

catlifeonmars — Tue, 28 Jul 2026 14:21:03 +0000

Most people on earth don’t make $200/mo.