HackerLangs
TopNewTrendsCommentsPastAskShowJobs

noperator

no profile record

Submissions

Pelican, or pelican't? A hint at Claude evals

noperator.dev
4 points·by noperator·28 hari yang lalu·0 comments

You can just say it

noperator.dev
3 points·by noperator·bulan lalu·0 comments

Show HN: Cagent – Agent in a Cage

github.com
2 points·by noperator·4 bulan yang lalu·0 comments

My AI got a GitHub account

maragu.dev
1 points·by noperator·6 bulan yang lalu·0 comments

Sift or Get Off the PoC: Vulnerability Research via Information Retrieval

arxiv.org
1 points·by noperator·6 bulan yang lalu·0 comments

Computing, Convivially

v5.chriskrycho.com
2 points·by noperator·6 bulan yang lalu·1 comments

comments

noperator
·7 hari yang lalu·discuss
Leanstral 1.5 has 6B active parameters. How many parameters does GPT-5.5 have?
noperator
·17 hari yang lalu·discuss
Yes, I did end up figuring out a clean way to allow normal reasoning inside <think> and then force JSON _after_ the closing </think>. Example here: https://gist.github.com/noperator/6c711ab19027ea8056442df839...
noperator
·18 hari yang lalu·discuss
Having some success while testing this model out as a replacement for GPT-5 nano in source code security review. Running on RTX 3090 (24 GB VRAM) via vLLM. It's not great on structured output (as noted in the model card) but I'm working around that in my harness.
noperator
·4 bulan yang lalu·discuss
> compute isolation means nothing if the sandbox can freely phone home.

Here's a project I've been working on to address the network risk. Uses nftables firewall allowing outbound traffic only to an explicit pinned domain allowlist (continuously refreshes DNS resolutions in the background).

https://github.com/noperator/cagent
noperator
·6 bulan yang lalu·discuss
Digital technologies can be terribly destructive, but they can also be gloriously empowering. How do we shape them into tools for conviviality rather than means of domination?