Hacker News: GTP

New comment by GTP in "GPT-2: Too Dangerous To Release (2019)"

GTP — Tue, 09 Jun 2026 21:25:22 +0000

> You know what the problem is in software engineering? A LOT of people have no clue what good software engineering is.

Indeed. Is Mythos going to change this?

New comment by GTP in "GPT-2: Too Dangerous To Release (2019)"

GTP — Tue, 09 Jun 2026 21:24:04 +0000

Has it been released to the public yet? Genuine question. Because if you didn't try it yourself, you have to rely on others' reports. And different people who tried it on different projects got different results, leading to different conclusions.

New comment by GTP in "The 29th International Obfuscated C Code Contest (IOCCC) 2025 Winners"

GTP — Sun, 07 Jun 2026 18:06:39 +0000

But, in the second case, you don't need permission. This is the crucial point.

New comment by GTP in "Show HN: Tired of duct-taping access control into agent prompts. Here's the fix"

GTP — Sun, 07 Jun 2026 18:05:10 +0000

in my case, we're using Docker containers to spwan Hermes instances, so it is easier to define what an instance is. And we need a container per user, because, without modifications, if multiple users talk to the same Hermes instance then they can access each other's conversations by asking the agent.

New comment by GTP in "The 29th International Obfuscated C Code Contest (IOCCC) 2025 Winners"

GTP — Sun, 07 Jun 2026 12:22:13 +0000

I think what they mean is that this sort of competition makes sense because it's about humans competing against each other, so that, even if we could have LLMs do it, the human version is still what captures our interest. In a similar way, we don't look at chess tournaments with computers playing against each others, but we look at chess grandmaster challenging each other. Even if it has been decades now that computers can beat grandmaster.

New comment by GTP in "The 29th International Obfuscated C Code Contest (IOCCC) 2025 Winners"

GTP — Sun, 07 Jun 2026 12:11:52 +0000

I don't think rule 7 would be self-contradictory since you indeed don't own the output of an LLM, but crucially, also no one else owns it. I read that rule as don't submit someone else's code without permission, which isn't violated by using an LLM.

The long tradition refers to the use of tooling in general, and could mean that, since past tools were accepted, recent tools like LLMs can be fair game as well.

But, since there can be doubts about this interpretation, them saying explicitly if LLMs are permitted or not could be beneficial. But then again, maybe they don't want to commit to an hard rule and have more freedom to decide on a case by case basis, or just don't advertise that LLMs are welcome to prevent a flood of vibe-coded submissions.

New comment by GTP in "Nvidia is proposing a beast of a CPU system for Windows PCs"

GTP — Sat, 06 Jun 2026 22:22:04 +0000

While I'm a supporter of Rust, I have to point out that Rust's memory safety doesn't help against side-channel attacks.

New comment by GTP in "Show HN: Tired of duct-taping access control into agent prompts. Here's the fix"

GTP — Thu, 04 Jun 2026 14:46:34 +0000

Maybe I'm not getting it either, but it looks like you're imagining a scenario where you deploy multiple instances of the agent, one per user, and then each one gets a different access level based on a key. Is this correct?

New comment by GTP in "Show HN: Tired of duct-taping access control into agent prompts. Here's the fix"

GTP — Thu, 04 Jun 2026 14:42:21 +0000

> The session sharing is something you've layered on top, with extra software

Nope. If you give access to the same agent instance to multiple people, then one user can get access to someone else's data. I'm working right now for a startup facing this issue. The easiest solution is to have a separate agent instance per user, but this increases resource consumption vs having a single agent instance with proper separation between multiple users.

Commodore Can't Survive Solely on Nostalgia

GTP — Wed, 03 Jun 2026 21:17:10 +0000

Article URL: https://www.timeextension.com/news/2026/06/commodore-cant-survive-solely-on-nostalgia-c64-ultimate-creator-will-announce-its-next-product-later-this-month

Comments URL: https://news.ycombinator.com/item?id=48390201

Points: 3

# Comments: 0

New comment by GTP in "A Gentle Introduction to Lattice-Based Cryptography [pdf]"

GTP — Sun, 31 May 2026 19:39:45 +0000

However, the time needed to get one plays a crucial role. Governments need to protect some piece of data for a very long time, but common people are generally fine with keeping something secret for their lives' duration. I don't care if someone decrypts my laptop's SSD after I'm dead.

New comment by GTP in "A Gentle Introduction to Lattice-Based Cryptography [pdf]"

GTP — Sun, 31 May 2026 12:58:42 +0000

Yes, but it exists because it was deemed better to be cautious and implement PQC despite the uncertainty and different points of view around the time scale to have cryptographically relevant quantum computers (or, from a different point of view, precisely due to the uncertainties). Their comment was in the wrong tone, but the doubts are there. BTW, PQC can be interesting to learn regardless of the discussion around quantum computers.

New comment by GTP in "Shift will clean homes for free to train future robots"

GTP — Sat, 30 May 2026 15:39:18 +0000

We need a captcha that differentiates between the two. I want to allow only good bots in my home.

New comment by GTP in "Microsoft BitLocker – YellowKey zero-day exploit"

GTP — Fri, 29 May 2026 09:31:15 +0000

Yes, but my question was where the files on the thumb drive are coming from. At first, my understanding was that you needed to get them from the victim's machine. But, after watching a LowLevel video on YouTube, it seems those are universal files that you can get from the exploit's repo.

New comment by GTP in "How do you build a semiconductor company on something that's free?"

GTP — Tue, 26 May 2026 21:14:13 +0000

> On the other hand, having it public makes exploits more likely since everyone can take a look.

Everyone can take a look, but not everyone will spend money to produce their own. So this will improve security, not reduce it.

New comment by GTP in "The user is visibly frustrated"

GTP — Tue, 26 May 2026 14:23:13 +0000

Plot twist: it opened a Moltbook account and leaked all your API keys :D

New comment by GTP in "Why is Vivado 2026.1 dropping Linux support for free tier?"

GTP — Sun, 24 May 2026 14:53:20 +0000

Some replies complain about the moderator not answering the main question. While this is a valid complaint, it is also likely that they don't know the answer as well. Now, the best reply would be to openly say that they don't have visibility into upper management's decisions. But, at the same time, I think it's possible that the way they're replying has to do with some internal guidelines on how to handle this sort of questions.

Microsoft's 6502 BASIC is now Open Source (2025)

GTP — Sun, 24 May 2026 13:16:33 +0000

Article URL: https://opensource.microsoft.com/blog/2025/09/03/microsoft-open-source-historic-6502-basic/

Comments URL: https://news.ycombinator.com/item?id=48257058

Points: 92

# Comments: 29

New comment by GTP in "A new book on Steve Jobs at NeXT"

GTP — Fri, 15 May 2026 18:07:49 +0000

I came here to comment on this as well, but from a different angle. Not only is the description inaccurate, but I distinctly remember a fellow HN commenter writing here years ago a very different story. IIRC, they claimed to be in Mac OS X's team. They said that, at the time, Jobs explicitly told them to not use Object Oriented Programming. But, since they knew he wouldn't be able to tell anyway, they still used OOP.

New comment by GTP in "Microsoft BitLocker – YellowKey zero-day exploit"

GTP — Fri, 15 May 2026 08:55:52 +0000

My concern isn't where they are copied to, but where they are copied from.