Hacker News: ra

New comment by ra in "Memdoor: A team of 5 Qwen3.8-27B models beats Fable 5 on LiveCodeBench Hard"

ra — Sun, 01 Nov 2026 01:03:04 +0000

"think", "believe" - maybe, but this is why we need better measurability.

New comment by ra in "Rune 1.3: adds TypeScript, a Helix editor, Kitty graphics and OSC-7501"

ra — Fri, 23 Oct 2026 22:57:29 +0000

This looks really cool - but are we still doing TUIs?

New comment by ra in "Superlite: A Faster Alternatie to Superpowers"

ra — Mon, 19 Oct 2026 04:48:52 +0000

Looking at the comments here it's clear that there's a lack of understanding about what skills fundamentally are, they're process or procedure.

The latest frontier models exhibit good process that was trained into them during the RL phase, and whilst that might make them massively more capable in terms of one-shot, or general coding ability, models without skills simply don't have a durable process.

If you have a need for procedure in your use of agents, you absolutely need skills like this. I'm not commenting on the relative merits of this or superpowers, but rather the seemingly religious debate, "skills are good" vs "skills are bad" discussion that keeps coming up.

New comment by ra in "DeepSeek search and fetch – $0.75 per 1000 successful requests"

ra — Sun, 18 Oct 2026 07:28:11 +0000

They're about half a gen behind. Very impressive to get this kind of utility from a flash series model... the impact of RL seems to be behind much of the coding and agentic improvements lately.

New comment by ra in "Ultramarine 1 Flash"

ra — Sat, 17 Oct 2026 03:38:16 +0000

Nice - I use GLM 5.3 flash for knowledge work now - will check this out.

New comment by ra in "Penguin Mail – open-source Rust email client for Linux with AI"

ra — Wed, 07 Oct 2026 06:03:19 +0000

Especially now these scam videos can be made on an industrial scale.

New comment by ra in "Powerless F1 drivers frustrated by Bahrain F1 software glitch"

ra — Mon, 05 Oct 2026 11:20:44 +0000

it has to be parameters only

New comment by ra in "Identity Management for Agentic AI [pdf] (2025)"

ra — Thu, 01 Oct 2026 22:19:47 +0000

I agree. If an agent needs to assume an identity, it needs a human in the loop to authorise - browser oauth is convenient for this.

Agents are not identities, they need to be granted permission to accomplish tasks. The identity is the person that grants the permission.

If no identity is required then this is simple machine-machine API keys.

New comment by ra in "The top secret URSALA, RAQUEL, and FARRAH satellites (2025)"

ra — Thu, 01 Oct 2026 07:05:09 +0000

I often see Hubble in the twilight, it's one of the few big satellites you can readily see with the naked eye - not quite as bright as the ISS but pretty close.

Why have i never seen these, is it simply that I wasn't looking for them?

New comment by ra in "OpenAI breaches Medicare, Albanese reveals"

ra — Thu, 24 Sep 2026 05:31:07 +0000

> no doubt built by an offshore contractor

More likely by a big 4 firm who collected fees exceeding AUD 100m

New comment by ra in "Show HN: Capsule – Single-file web apps that save their data into SQLite"

ra — Wed, 16 Sep 2026 08:57:11 +0000

Ah yes that was a whole era of programming for me. It was really fast to ship features... and then came the web!

New comment by ra in "EU Floats Canada Becoming the Bloc's First 'Associate Member'"

ra — Wed, 16 Sep 2026 08:47:57 +0000

I think the world would be jumping into alignments like this if the US hadn't spent the last 80 years building a "nuclear umbrella" for the "western alliance" with itself in the centre.

I bet many countries would prefer to disassociate with US foreign policy right now if they weren't security dependent.

I expect after this destructive episode in history has played out, if we're all still alive and free in another decade then there will be alternative umbrellas.

New comment by ra in "Dramatic insider warnings over AI fall flat with some in Silicon Valley"

ra — Sun, 13 Sep 2026 09:31:45 +0000

Greetings professor Falcon.

New comment by ra in "OpenAI Agents API"

ra — Sun, 13 Sep 2026 03:42:17 +0000

The best advice is to use evals to improve any code that produces output that goes into the LLMs context.

New comment by ra in "I spent $220 on Google app ads and 60% of the installs were robots"

ra — Sat, 12 Sep 2026 06:18:20 +0000

I imagine it's a side-effect of another fraud prevention effort. In my experience ad fraud on a google search ad is at least 17% of traffic, and potentially a lot more.

Whilst I don't doubt that Google invest heavily in fraud prevention, they're not actually succeeding.

New comment by ra in "Google will buy half the electricity from one of Finland's nuclear power plants"

ra — Fri, 11 Sep 2026 05:20:35 +0000

We will all be happy with our local inference, but how will our overloards be able to surveil and control us en-masse without them?

New comment by ra in "OpenAI Agents API"

ra — Thu, 10 Sep 2026 23:51:13 +0000

> Async tool calls, having the agent wait indefinitely for a human response, and showing a form or questions to the user via a tool call are a few common capabilities that come up that a product manager might miss at first.

All of this is specified in the ACP spec, so if you build your agents from that - you don't end up skipping features.

Also vital is proper prompt caching, tool design and some connection retry mechanism.

New comment by ra in "A mysterious kidney disease has arrived in Texas"

ra — Fri, 04 Sep 2026 06:18:35 +0000

except maybe Japan?

New comment by ra in "JupyterGIS 0.16: a grammar of graphics for maps, and collaborative story maps"

ra — Sun, 30 Aug 2026 11:12:52 +0000

That's quite impressive!

New comment by ra in "Iceland rejects reopening talks on EU entry"

ra — Sun, 30 Aug 2026 09:52:17 +0000

that's also pretty much what happened in Brexit