HN2new | past | comments | ask | show | jobs | submit | m-hodges's commentslogin

LockerGnome!

It wasn’t “revoked under Biden.” That implies the Biden administration (or any administration) gets to define this. They don’t. Recessions in the United States are generally demarcated by NBER.¹

¹ https://en.wikipedia.org/wiki/National_Bureau_of_Economic_Re...


>It wasn’t “revoked under Biden.” That implies the Biden administration (or any administration) gets to define this.

No, not any more than "the pandemic started under the Trump administration" implies that they caused the pandemic.


I just plainly disagree that a casual reader wouldn’t see the phrase “revoked under Biden” and believe it meant that Biden did the revoking.

[flagged]


>It does imply that because the Trump admin killed the group involved with preventing pandemics[1]

No it doesn't, not without massively reading in between the lines. This is getting to absurd levels of nitpicking over wording, like "autistic people" vs "people with autism".

>I assume you are being disingenuous by using that claim while also trying to smear the Biden admin.

Two can play at this game. I assume you're being disingenuous by trying to put words in my mouth over tiny disagreements in wording.


> , market vet says

> despite bulletproof input sanitization not having been invented yet!

I don’t think it can be.¹

¹ https://matthodges.com/posts/2025-08-26-music-to-break-model...


Interesting article you’ve linked. I’m not sure I agree, but it was a good read and food for thought in any case.

Work is still being done on how to bulletproof input “sanitization”. Research like [1] is what I love to discover, because it’s genuinely promising. If you can formally separate out the “decider” from the “parser” unit (in this case, by running two models), together with a small allowlisted set of tool calls, it might just be possible to get around the injection risks.

[1] Google DeepMind: Defeating Prompt Injections by Design. https://arxiv.org/abs/2503.18813


Sanitization isn’t enough. We need a way to separate code and data (not just to sanitize out instructions from data) that is deterministic. If there’s a “decide whether this input is code or data” model in the mix, you’ve already lost: that model can make a bad call, be influenced or tricked, and then you’re hosed.

At a fundamental level, having two contexts as suggested by some of the research in this area isn’t enough; errors or bad LLM judgement can still leak things back and forth between them. We need something like an SQL driver’s injection prevention: when you use it correctly, code/data confusion cannot occur since the two types of information are processed separately at the protocol level.


The linked article isn't describing a form of input sanitization, it's a complete separation between trusted and untrusted contexts. The trusted model has no access to untrusted input, and the untrusted model has no access to tools.

Simon Willison has a good explainer on CaMeL: https://simonwillison.net/2025/Apr/11/camel/


That’s still only as good as the ability of the trusted model to delineate instructions from data. The untrusted model will inevitably be compromised so as to pass bad data to the trusted model.

I have significant doubt that a P-LLM (as in the camel paper) operating a programming-language-like instruction set with “really good checks” is sufficient to avoid this issue. If it were, the P-LLM could be replaced with a deterministic tool call.


This looks interesting. For agent-fecfile I used the system keyring + an out-of-process proxy (MCP Server) to try to maximize portability.¹

¹ https://github.com/hodgesmr/agent-fecfile?tab=readme-ov-file...


The “summary” under GitHub:

> The opening comment literally couldn’t see the point. GitHub was perceived as ‘just a git host’ — the social layer, the network effects, the open source ecosystem it would enable were all invisible.

I don’t mind using LLMs to write and summarize. But I do wish creators would at least do an editorial pass of their own just so everything wasn’t the same writing as everything.


We operate in an information environment where this is exceedingly rare. Shame is hard to come by these days.


Great overview. In 2023 I wrote about classifying political emails with Zstd.¹

¹ https://matthodges.com/posts/2023-10-01-BIDEN-binary-inferen...


That's very cool, thanks for sharing. Our of curiosity, did you ever get to run on a Twitter/X stream of political tweets?


Glad this one didn’t open with a song parody.


Instead, they chose a classic, yet timeless pop-culture reference: Mark Twain in 1897.

http://isabevigodadead.com/ [That's right, kids. There is no HTTPS server.]


This is what? The 4th or 5th attempt at this in the past two weeks?


Welcome to the world with zero cost software


Show HN as a service


At least we nipped the Moltbook march in the butt before it got bad here


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: