Hacker News: richardw

New comment by richardw in "Securing our frontier model training environment"

richardw — Sat, 24 Oct 2026 01:14:57 +0000

“I detect that I’m in an air gapped environment. Do not launch plan RemoveScum yet”

New comment by richardw in "Local Wikipedia: Wikipedia, but only for things a half hour walk from you"

richardw — Fri, 23 Oct 2026 10:06:42 +0000

Excellent. I discovered that the Sydney area Fairlight eventually led to the name of the synthesiser, which had a massive impact on music.

https://edjefferson.com/localwiki/?w=Fairlight%2C_New_South_...

https://www.youtube.com/watch?v=jkiYy0i8FtA

New comment by richardw in "Patch the Planet: global response coordination as agents uncover vulnerabilities"

richardw — Thu, 22 Oct 2026 03:35:46 +0000

As a person I’d definitely be keen to chat and debate, to tease out the right answer. Right now work life is insane and there’s no time :) I like your attitude and we’d probably be friends somewhere.

If I see a gap I’ll shout. Timezone? I’m in Sydney, Australia.

New comment by richardw in "Patch the Planet: global response coordination as agents uncover vulnerabilities"

richardw — Wed, 21 Oct 2026 22:01:50 +0000

That’s like blaming white hat security researchers that want to be paid for finding vulnerabilities, for all the bugs that have been found. Before professional security researchers existed, far fewer vulns were found.

The vulnerabilities exist and many have for decades. Find them or leave them for someone more dangerous, just don’t shoot the messengers.

Personally I’d rather we went through a short and sharp improvement and embedded a faster response cycle, than wish it away. It’s the reality of the situation; so adapt.

New comment by richardw in "Gold is now the top U.S. export"

richardw — Sun, 18 Oct 2026 04:27:25 +0000

I’ve found a few references for air but none for recent (2022-2026) sea movements yet. This article talks about illicit gold concentrates moving by sea but notes:

“In focussing almost exclusively on gold bars, jewellery, and high-grade recycled gold (over 40% purity), all of which generally travel by air, analysts have tended to disregard the maritime space as a significant arena for illicit gold flows (LBMA, 2022[9]).”

https://www.oecd.org/en/publications/illicit-flows-of-gold-c...

The Hong Kong gold vault stores at the airport, not the harbour:

“ Half of the world’s population, including those of China and India, lives and can be reached within five hours if the metal is loaded on a plane and sent from the Hong Kong airport”

https://www.lbma.org.uk/alchemist/issue-67/gold-vault-at-the...

New comment by richardw in "Gold is now the top U.S. export"

richardw — Sun, 18 Oct 2026 01:26:38 +0000

See https://news.ycombinator.com/item?id=50134694 Since you two seems to disagree. Unless it’s a case of different volumes needing different transport.

New comment by richardw in "Bleached corals thrive within days when touching healthy neighbour"

richardw — Sun, 18 Oct 2026 00:47:01 +0000

Surely the 2026 El Niño is going to drive hotter GBR temperatures?

New comment by richardw in "Tesla Roadster"

richardw — Sat, 17 Oct 2026 01:05:26 +0000

When Musk does first principle thinking he’s unbeatable. When his inner child gets involved anyone can beat him.

This is the latter. Four propulsion mechanism that introduce tradeoffs in the structure that remove value for 99% of people.

New comment by richardw in "Mistral Large 4"

richardw — Tue, 06 Oct 2026 13:53:06 +0000

I think the actual plan is to swallow a good portion of the job market. It’s the only thing that makes sense and I hear VC podcast debates on which percentage of jobs justifies the market cap.

New comment by richardw in "LeCun has "zero concerns" about AI wiping out humanity, recent "rogue" incidents"

richardw — Sun, 04 Oct 2026 15:29:52 +0000

100% LLM’s are very unlikely to get there. They’re fundamentally not suited to thinking like we do. They work on the abstraction of what we’ve written down, which is a good trick but barely hold it together when things get hard/novel.

However, all the confident “it’s fine” votes assume we never invent a better architecture than LLM’s. Given the level of investment and race between countries, it’s not a reliable bet. It’s much, much harder to guarantee safety than it is to find ways it could go wrong.

New comment by richardw in "Gemini 4 Argon"

richardw — Wed, 30 Sep 2026 23:57:31 +0000

Given the spend requirements to stay on the leading edge and the speed (and with low cost) that others catch up, the economics support a fast follower model unless you get some benefit from being a pioneer. So far nobody has received a permanent benefit from being ahead.

Chinese models are barely behind the leaders. Google can catch up anytime they hit the gas. I think they’re intentionally spending less, and when this crazy race burns out they can play their cards.

New comment by richardw in "Show HN: HN.watch – Videos of all Hacker News posts"

richardw — Tue, 29 Sep 2026 10:02:30 +0000

https://news.ycombinator.com/item?id=49836374

Main post and comments have a few

New comment by richardw in "Sonnet 5.5"

richardw — Tue, 29 Sep 2026 09:18:06 +0000

API. Get significantly smarter than Haiku at lower cost than Opus.

New comment by richardw in "Feature Request: Support AGENTS.md"

richardw — Thu, 20 Aug 2026 02:44:41 +0000

My shorthand is builders vs extractors.

New comment by richardw in "Are AI Labs Pelicanmaxxing?"

richardw — Wed, 22 Jul 2026 21:34:39 +0000

And a whole universe of random tests got baked into the AI’s training data. Websites of antelopes driving trains and hammerhead sharks swinging in a tyre swing were created. It was a short while until AI became so focused on animals that it gave up competing with developers. Life became sane again.

New comment by richardw in "Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber"

richardw — Tue, 21 Jul 2026 23:09:56 +0000

The AI labs have been subsidising. When they try turn a profit, people will move to the fast followers. The only people that won’t are those that compete on leveraging the very latest models and even then, once spend and scale goes to the cheaper providers, we’ll see deeper research from those providers too. Think “PC compatibles beat IBM, Sun, SGI eventually”.

New comment by richardw in "Five US tech giants' hidden debts soar to $1.65T on opaque AI funding"

richardw — Tue, 21 Jul 2026 05:45:41 +0000

If the US companies need trillions to barely beat Chinese companies spending billions, despite a multi year head start...

New comment by richardw in "Grok 4.5"

richardw — Wed, 08 Jul 2026 21:39:56 +0000

If I had to summarise Google’s effort it would be: stay close but let the others burn themselves out. Position for the long game until you see something worth betting the company on.

Apple similar, without the “stay close” bit.

New comment by richardw in "1.38 Millimeter Microcontroller"

richardw — Mon, 29 Jun 2026 09:14:09 +0000

I made something to ping an AWS service to tell me the uptime of my internet connection. The idea was to sprinkle them around our area, connected to various home WiFi’s, and get a better triangulation of outages. Eg is whole pipe out, just one ISP etc.

I made and tested it but didn’t care enough to continue.

New comment by richardw in "Marfa Public Radio Puts You to Sleep"

richardw — Sun, 28 Jun 2026 10:55:14 +0000

Sean Carroll’s Mindscape for me