Hacker News: erwald

New comment by erwald in "The LLM Critics Are Right. I Use LLMs Anyway"

erwald — Thu, 16 Jul 2026 12:52:39 +0000

Master craftsmen paid apprentices almost next to nothing, and they were often contractually guaranteed to stick around for many years, so the teaching was a kind of wage and also a cost that could be recuperated later on. (The apprentice even often had to pay the craftsman to take them on.) None of those things are true for junior software engineers, who are paid to contribute and can leave at any moment. Also, yes apprentices often had to do chores. It is just not analogous at all.

New comment by erwald in "AI 2040: Plan A"

erwald — Sat, 11 Jul 2026 14:59:02 +0000

No, 2027 was never their median forecast; it was their modal forecast. See https://blog.aifutures.org/p/clarifying-how-our-ai-timelines...

It's more like it moved from 2028-2032 to 2030-2035 (depending on the author).

New comment by erwald in "China will likely have its own Mythos-like model around February 2027"

erwald — Tue, 07 Jul 2026 13:04:52 +0000

I think electricity doesn't matter that much (yet) because China is bottlenecked on chips. I think the incentives/directives to build on Huawei also doesn't matter that much yet because it's still such a small percentage of compute relative to NVIDIA, even for Chinese AI companies. (But this too could matter more from 2027-2030 and on.)

New comment by erwald in "Austria Lobbies EU to Host Anthropic After US Access Curbs"

erwald — Wed, 01 Jul 2026 15:56:14 +0000

Yeah, to be clear I'm pretty excited about confidential computing and startups building on it, like Tinfoil, for some use cases. I just wanted to point out it's far from adequate for some important threat models (e.g., securing model weights for data centers located abroad, I think). (It's also not super widely adopted in AI yet, but that seems to be changing, at least for inference workloads.)

New comment by erwald in "Austria Lobbies EU to Host Anthropic After US Access Curbs"

erwald — Wed, 01 Jul 2026 13:25:59 +0000

Where did you get this information? I think it's wrong -- I'm pretty sure they used the Export Administration Regulations (EAR) under the Export Control Reform Act (ECRA), which is under Commerce, not ITAR which is under the State Department. See for example https://harvardlawreview.org/blog/2026/06/is-access-to-fable...

New comment by erwald in "Austria Lobbies EU to Host Anthropic After US Access Curbs"

erwald — Sun, 28 Jun 2026 15:37:34 +0000

Confidential computing is not secure against a potential attacker who has physical access to the hardware. The CC security guarantees explicitly assume the attacker has no physical access.

New comment by erwald in "Austria Lobbies EU to Host Anthropic After US Access Curbs"

erwald — Sun, 28 Jun 2026 15:36:12 +0000

I don't think ITAR has anything to do with any of this.

New comment by erwald in "Project Glasswing: An Initial Update"

erwald — Sat, 23 May 2026 11:57:06 +0000

What are your general, vibes-based impressions of Mythos so far?

Women should be able to open things

erwald — Fri, 22 May 2026 18:07:56 +0000

Article URL: https://worldspiritsockpuppet.com/2026/05/04/women-open-things.html

Comments URL: https://news.ycombinator.com/item?id=48239288

Points: 3

# Comments: 7

How much US compute is China renting from the cloud?

erwald — Fri, 22 May 2026 18:03:40 +0000

Article URL: https://www.the-substrate.net/p/how-much-us-compute-is-china-renting

Comments URL: https://news.ycombinator.com/item?id=48239244

Points: 2

# Comments: 0

New comment by erwald in "Nobel laureate Olga Tokarczuk used AI while writing her latest novel"

erwald — Wed, 20 May 2026 15:31:54 +0000

Seems pretty reasonable!

How much should we worry about secretly loyal AIs?

erwald — Wed, 20 May 2026 15:28:29 +0000

Article URL: https://www.the-substrate.net/p/how-much-should-we-worry-about-secretly

Comments URL: https://news.ycombinator.com/item?id=48209360

Points: 3

# Comments: 1

How banned AI chips end up in China

erwald — Fri, 15 May 2026 16:22:57 +0000

Article URL: https://www.the-substrate.net/p/how-banned-ai-chips-end-up-in-china

Comments URL: https://news.ycombinator.com/item?id=48150524

Points: 4

# Comments: 1

Securing AI infrastructure to prevent backdoors and sabotage

erwald — Sun, 05 Apr 2026 19:47:02 +0000

Article URL: https://www.the-substrate.net/p/securing-ai-infrastructure-to-prevent

Comments URL: https://news.ycombinator.com/item?id=47653133

Points: 2

# Comments: 0

3k languages are dying, but more are being invented

erwald — Tue, 10 Mar 2026 12:21:16 +0000

Article URL: https://asteriskmag.substack.com/p/3000-languages-are-dying-but-more

Comments URL: https://news.ycombinator.com/item?id=47322236

Points: 3

# Comments: 1

OpenAI has released Dow contract language, and it's as Anthropic claimed

erwald — Sun, 01 Mar 2026 01:05:46 +0000

Article URL: https://twitter.com/justanotherlaw/status/2027855993921802484

Comments URL: https://news.ycombinator.com/item?id=47202477

Points: 3

# Comments: 1

Where will China get its compute in 2026?

erwald — Fri, 13 Feb 2026 20:54:16 +0000

Article URL: https://www.the-substrate.net/p/where-will-china-get-its-compute

Comments URL: https://news.ycombinator.com/item?id=47007679

Points: 1

# Comments: 0

New comment by erwald in "GLM-5: Targeting complex systems engineering and long-horizon agentic tasks"

erwald — Thu, 12 Feb 2026 00:06:32 +0000

Thanks. I'm like 95% sure that you're wrong, and that GLM-5 was trained on NVIDIA GPUs, or at least not on Huawei Ascends.

As I wrote in another comment, I think so for a few reasons:

1. The z.ai blog post says GML-5 is compatible with Ascends for inference, without mentioning training -- it says they support "deploying GLM-5 on non-NVIDIA chips, including Huawei Ascend, Moore Threads, Cambricon, Kunlun Chip, MetaX, Enflame, and Hygon" -- many different domestic chips. Note "deploying". https://z.ai/blog/glm-5

2. The SCMP piece you linked just says: "Huawei’s Ascend chips have proven effective at training smaller models like Zhipu’s GLM-Image, but their efficacy for training the company’s flagship series of large language models, such as the next-generation GLM-5, was still to be determined, according to a person familiar with the matter."

3. You're right that z.ai trained a small image model on Ascends. They made a big fuss about it too. If they had trained GLM-5 with Ascends, they likely would've shouted it from the rooftops. https://www.theregister.com/2026/01/15/zhipu_glm_image_huawe...

4. Ascends just aren't that good

New comment by erwald in "GLM-5: Targeting complex systems engineering and long-horizon agentic tasks"

erwald — Wed, 11 Feb 2026 23:56:17 +0000

Kudos for changing your mind

New comment by erwald in "GLM-5: Targeting complex systems engineering and long-horizon agentic tasks"

erwald — Wed, 11 Feb 2026 23:37:11 +0000

Thanks. I'm like 95% sure that you're wrong (as is the parent), and that GLM-5 was trained on NVIDIA GPUs, or at least not on Huawei Ascends.

I think so for a few reasons:

1. The Reuters article does explicitly say the model is compatible with domestic chips for inference, without mentioning training. I agree that the Reuters passage is a bit confusing, but I think they mean it was developed to be compatible with Ascends (and other chips) for inference, after it had been trained.

2. The z.ai blog post says it's compatible with Ascends for inference, without mentioning training, consistent with the Reuters report https://z.ai/blog/glm-5

3. When z.ai trained a small image model on Ascends, they made a big fuss about it. If they had trained GLM-5 with Ascends, they likely would've shouted it from the rooftops.

4. Ascends just aren't that good

Also, you can definitely train a model on one chip and then support inference on other chips; the official z.ai blog post says GLM-5 supports "deploying GLM-5 on non-NVIDIA chips, including Huawei Ascend, Moore Threads, Cambricon, Kunlun Chip, MetaX, Enflame, and Hygon" -- many different domestic chips. Note "deploying".