Hacker News: nojs

New comment by nojs in "Improving GPT‑5.6 Sol in ChatGPT, expanding GPT‑5.6 Luna access for free users"

nojs — Thu, 06 Aug 2026 22:44:46 +0000

Auto-effort and similarly auto model routing suffer from a halting problem sort of issue: you don’t reliably know if a request is complex unless you use a complex model to make the decision.

New comment by nojs in "AMD acquires Taalas to boost inference performance by etching models in silicon"

nojs — Thu, 06 Aug 2026 21:43:03 +0000

Can anyone comment on the economics and likely turnaround times of this process, when it’s more mature?

Would it be realistic for a frontier lab to deploy this or would the turnaround time mean the model is always too out of date?

Assuming the weights and architecture are eventually stable, how much cheaper would this end up being?

New comment by nojs in "TIME Is Serving AI Bots a Different Website, with Ads Built In"

nojs — Wed, 05 Aug 2026 20:28:37 +0000

I mean isn’t this just like feeding the Google crawler a different version of your website than humans, stuffed with keywords? One would assume the new breed of crawlers can deal with it similarly.

New comment by nojs in "Don't be a meat proxy"

nojs — Mon, 03 Aug 2026 07:35:05 +0000

The problem is a lot of people ask questions that claude et al can objectively answer way better than me, and they didn’t think to ask.

Should I a) give a worse answer b) ask claude and launder it as a meat proxy or c) tell them to ask claude?

b and c piss people off but often get them to a better answer.

New comment by nojs in "AI financial advice is surprisingly good, especially if you ask right questions"

nojs — Sun, 02 Aug 2026 07:04:22 +0000

For a minute I thought your comment was a lead in to this south park skit: https://youtu.be/-DT7bX-B1Mg

New comment by nojs in "How to Spot AI Writing"

nojs — Sat, 01 Aug 2026 14:58:08 +0000

The best resource i’ve seen on this is https://en.wikipedia.org/wiki/Wikipedia:Signs_of_AI_writing

New comment by nojs in "qm – Multiplayer agent harness for work"

nojs — Sat, 01 Aug 2026 00:02:29 +0000

> Em-dash (—) is COMPLETELY banned. It is the LLM's signature stylistic crutch and it is the #1 visual Tell in production tests. There is no "limited use" allowance, no "natural language frequency" allowance, no "in body copy is fine" allowance. None.

I guess the em dash is really dead.

New comment by nojs in "Anti-fraud tools can't keep pace with robocall scammers"

nojs — Fri, 31 Jul 2026 17:45:47 +0000

Perhaps the future of captcha is anti-captcha. What agent can resist responding with a rhyming couplet or inverting a binary tree?

New comment by nojs in "Investigating three real-world incidents in our cybersecurity evaluations"

nojs — Fri, 31 Jul 2026 01:44:37 +0000

> For instance, in one case, in order to create a PyPI account, Claude needed an email address. And in order to create an email address, it needed a phone number. To get a phone number, after failing to find a free phone number service, it tried—and failed—to obtain funds to pay for a phone number through several different means.

Makes you wonder about the next steps an overly tenacious agent might take to pursue an innocuous goal at all costs.

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

nojs — Thu, 30 Jul 2026 01:45:36 +0000

Flagged as LLM slop.

New comment by nojs in "Good CLI Design Is Mostly Silence"

nojs — Wed, 29 Jul 2026 09:22:57 +0000

Flagged as slop. What is the point of posting something like this if you can’t even be bothered to write it yourself?

New comment by nojs in "Interview with Boris Cherny [video]"

nojs — Wed, 29 Jul 2026 03:34:36 +0000

> From what I'm hearing, you for every model release, you basically delete all of the codebase, delete all of the prompt and start from scratch every time. That in the old world would have been not something Startup would have done for the product. It's like press delete every six months for everything.

> That's right. So to be fair, we don't delete the entire codebase.

New comment by nojs in "Discovering Cryptographic Weaknesses with Claude"

nojs — Wed, 29 Jul 2026 03:04:42 +0000

For one thing they are extremely GPU constrained due to export controls, so it’s unlikely a priority compared to training

New comment by nojs in "Kimi K3 Now Available via Telnyx Inference API"

nojs — Tue, 28 Jul 2026 08:29:24 +0000

This happens with other models too - Gemini often identifies as ChatGPT for me, confusing many a debugging attempt

New comment by nojs in "The relay market powering token resellers and fraud"

nojs — Sun, 26 Jul 2026 20:10:58 +0000

Seems to be an LLM megaexpansion of the actual source (in chinese): https://www.v2ex.com/t/1196011

New comment by nojs in "Open-weight AI is having its Kubernetes moment"

nojs — Sat, 25 Jul 2026 20:52:43 +0000

What hardware are you expecting to run K3 on?

New comment by nojs in "The case for MUDs in modern times (2018)"

nojs — Sat, 25 Jul 2026 01:57:02 +0000

FYI the wikipedia link in the footer is broken (points to mod, not mud)

New comment by nojs in "I tricked Claude into leaking your deepest, darkest secrets"

nojs — Wed, 15 Jul 2026 08:21:51 +0000

This is not about admin rights, it’s about the agent leaking information it knows from its memories. Sandboxing won’t really help you.

New comment by nojs in "Precursor"

nojs — Mon, 13 Jul 2026 15:55:07 +0000

I mean CF already forces 5 minutes of motorbike identification on anyone not in a whitelisted western country, so a small percentage of blind people is unlikely to worry them.

New comment by nojs in "Ask HN: Add flag for AI-generated articles"

nojs — Mon, 13 Jul 2026 03:45:39 +0000

Not OP, but I’ve been nuked with downvotes for this several times too and tend to delete the dead comments. The slop is so prevalent that at this point it’s not a particularly interesting thing to say I think.