Hacker News: novaleaf

New comment by novaleaf in "Auto mode is now the default in Claude Code"

novaleaf — Mon, 10 Aug 2026 14:12:19 +0000

> "Adversarial red-teaming made auto mode stronger"

Has to be said: they (and OpenAi) don't let you red-team your own creations, pisses me off to no end.

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

novaleaf — Sun, 09 Aug 2026 21:21:44 +0000

opus 5 doesn't follow instructions that great. probably needed that extra "creativity" to benchmax. if you are stuck on anthropic, try Opus 4.8 or Fable5 a try with the same prompts. very different results.

New comment by novaleaf in "AI-Generated Images Discourage Me from Reading Your Blog"

novaleaf — Tue, 04 Aug 2026 15:05:31 +0000

that's pretty nice illustration work, I'd agree, keep doing it :D

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

novaleaf — Tue, 28 Jul 2026 00:12:56 +0000

Frontier models can hack you, we should have access to tools assisting defense.

I ranted about this in a prior thread [1]

Claude doesn't have a "Security whitelist" for small biz. Codex does, but they never replied to my application. This is a great example why, as of today, everyone NEEDS access to the Open Weight models.

[1]: https://news.ycombinator.com/item?id=49035303#49040674

New comment by novaleaf in "It's not empowering to hand off the details"

novaleaf — Mon, 27 Jul 2026 02:28:58 +0000

as others mentioned, it sounds like "Off to Be the Wizard". Unfortunately I wouldn't call it a great series. It starts off strong but drags. IIRC I got through half of the second book before stopping.

New comment by novaleaf in "Nvidia, Microsoft, Meta warn against overregulating open-weight models"

novaleaf — Sun, 26 Jul 2026 14:34:50 +0000

With Openai and Anthropic, you can opt-out of training. Kimi you can't.

New comment by novaleaf in "The new rules of context engineering for Claude 5 generation models"

novaleaf — Sat, 25 Jul 2026 23:32:42 +0000

> I guess part of it is also that I don't mind doing 'hand-edits' like for example LLMs love to say "// so and so removed" I just go and remove that manually later rather than being like "don't comment about what you removed!11" cause you're really fighting deep grooves in the model's behavior at that point.

Maybe I'm misunderstanding you, but that's just about the best example possible for using AGENTS/CLAUDE.md. Just add "don't comment about what you removed!11" and you never have to say it again...

...but you'll get constantly nagged about the `11` of course!

New comment by novaleaf in "Nvidia, Microsoft, Meta warn against overregulating open-weight models"

novaleaf — Sat, 25 Jul 2026 17:52:55 +0000

Agree, out of the two, I can get Codex to design and implement security systems (Fable just refuses to discuss). I've only used K3 a little bit, but I found that being able to discuss attack vectors and their mechanics gives me details to paste back into Codex for it to ingest. I could never have gotten there through Codex alone.

New comment by novaleaf in "Nvidia, Microsoft, Meta warn against overregulating open-weight models"

novaleaf — Sat, 25 Jul 2026 15:57:25 +0000

Yeah it's not cheap, it's just the normal $200 subs for each.

New comment by novaleaf in "Nvidia, Microsoft, Meta warn against overregulating open-weight models"

novaleaf — Sat, 25 Jul 2026 14:53:25 +0000

I applied for individuals. I did their verification steps and answered questions in a few minutes, so that was indeed easy. The problem is that was all that happened. No followup, no access, no denial. When I tried to reapply it tells me I can't apply again.

OpenAi (and Anthropic) have no incentive to allow security access to individuals. I'm not a deep-pocket org or influential gov agency. Allowing individuals increases the risk of bad press (what if I do something naughty and talk about it?) so best to ignore us.

New comment by novaleaf in "Nvidia, Microsoft, Meta warn against overregulating open-weight models"

novaleaf — Sat, 25 Jul 2026 14:50:11 +0000

keep in mind that Kimi trains on your data, so if you care about that, sandbox what it has access to.

New comment by novaleaf in "Nvidia, Microsoft, Meta warn against overregulating open-weight models"

novaleaf — Fri, 24 Jul 2026 22:27:52 +0000

I think security is an integral part of any production software, and if you get value from LLM's in software development, it seems likely they can be useful in security too.

My point and frustration is that gatekeeping in the name of Security makes the Chinese models actually better at security than USA models.

New comment by novaleaf in "Nvidia, Microsoft, Meta warn against overregulating open-weight models"

novaleaf — Fri, 24 Jul 2026 20:19:21 +0000

I don't think it's nefarious, but the end result leads to a pretty frustrating experience by anybody needing actual security work. (and without the organizational deep pockets to obtain SOC 2 attestation)

New comment by novaleaf in "Nvidia, Microsoft, Meta warn against overregulating open-weight models"

novaleaf — Fri, 24 Jul 2026 19:39:20 +0000

Some irony: I subscribe to Claude and Codex (20x plans), and now Kimi.

Why Kimi? because K3 is the only frontier model I can have a serious conversation with about my product's security.

(I did apply for OpenAi's Cyber Pilot but got no response)

New comment by novaleaf in "OpenAI and Hugging Face address security incident during model evaluation"

novaleaf — Tue, 21 Jul 2026 21:59:42 +0000

Reminds me of the gain-of-function, COVID lab leak hypothesis. It seems like humanity just can't stay away from Pandora's box.

New comment by novaleaf in "Kimi K3: Open Frontier Intelligence"

novaleaf — Thu, 16 Jul 2026 18:28:12 +0000

frontier vs "not quite" :D

New comment by novaleaf in "Grok 4.5"

novaleaf — Fri, 10 Jul 2026 16:33:54 +0000

makes me wonder if the starbucks story is fiction too :P

New comment by novaleaf in "GPT-5.6"

novaleaf — Fri, 10 Jul 2026 16:33:16 +0000

no, I didn't get access to sol until a few hours ago. I just have my claude protocol files linked to inside codex. trying Sol this morning for the first time, so I can't really comment on that vs gpt5.5.

However you can do what you are asking "fable--> sol" you need to setup a mcp or have fable run a bash tool, just invoke the `codex.exe` cli tool with whatever cmdline args are needed.

New comment by novaleaf in "GPT-5.6"

novaleaf — Thu, 09 Jul 2026 17:36:36 +0000

I sub both codex and claude at 20x. I like opus+fable more than gpt5.5 because it seems gpt tries to finish tasks by leaving any ambiguity unresolved. claude seems better at surfacing open questions.

This is using the same AGENTS.md prompts, which were designed firstly for Claude use, so maybe it's something that could be optimized better if I understood gpt as well?

New comment by novaleaf in "Grok 4.5"

novaleaf — Wed, 08 Jul 2026 19:59:59 +0000

I think the story was Starbucks -> Seattle's Best Coffee, not McD's --> BK. but it does work I guess.