Hacker News: dada216

New comment by dada216 in "GLM Built Its Own Inference Infrastructure"

dada216 — Thu, 17 Sep 2026 10:07:18 +0000

We built a complete production-grade inference service from scratch on a cluster of more than 100,000 Chinese-made AI accelerators. All production inference for GLM-5.3-Flash runs on this system.

New comment by dada216 in "Open-source AI and open models reading list"

dada216 — Mon, 14 Sep 2026 11:11:51 +0000

the fact that it's mostly about policies does not make the link list useless.

New comment by dada216 in "Ask HN: Anyone still coding like 2021? Where do you work?"

dada216 — Wed, 09 Sep 2026 14:38:59 +0000

What kind of programming do you do? If it's enterprise programming and it's basically CRUDs all the way down, yeah, LLM are pretty good at it and you need to find a balance if you don't enjoy it anymore.

All this bitching does feel a little off to me though, I get it, but I don't think you guys get how bad it sounds from the outside.

LLMs are not perfect, even in enterprise programming. Even in my case, where a few CRUDs may manage 100s of millions of dollars, and the code LLMs produce is actually good enough, for sure comparable with mid developers, most of them, you do need to develop a system to keep everything on track, I've found my balance and honestly? It's good and I enjoy it.

Go into embedded critical systems, that's probably the last thing they will get good at.

New comment by dada216 in "The unbearable cheapness of open weight models"

dada216 — Fri, 26 Jun 2026 09:00:06 +0000

Then they will leave the huge advantage in cost to the competition, I mean their customers competitors. Hard to fathom how US companies will not want to use the cheaper option when EU and Asian companies can.

New comment by dada216 in "The unbearable cheapness of open weight models"

dada216 — Fri, 26 Jun 2026 08:56:45 +0000

I have deployed very successful LLM based software that reads sales people emails and inserts orders, or stuff akin to orders, in the rest of the systems. Can you write me a regex that parses a messy human email thread and produces a clean JSON with all the order details? It's been working for a year with less than 3% error rate, better than what the humans themselves were doing.

New comment by dada216 in "No-One Escapes the Permanent Underclass"

dada216 — Fri, 26 Jun 2026 06:22:49 +0000

I remember an episode of TNG were Riker was instructing the computer to come up with an algorithm to search the surrounding space with certain sensor in a specific pattern, I don't remember the exact words but it was something like a sphere with growing radius with the enterprise at the center. we now, basically, have that, except for the space ship.

New comment by dada216 in "Volkswagen started blocking GrapheneOS users"

dada216 — Wed, 17 Jun 2026 15:49:14 +0000

Those two set of systems are separate and very distinct.

New comment by dada216 in "Ask HN: Has anyone replaced Claude/GPT with a local model for daily coding?"

dada216 — Mon, 15 Jun 2026 17:19:31 +0000

Local? No. Via opencode Go subscription using GLM mainly? Yes, I still use Gemini/Claude/GPT via api from openrouter for adjacent tasks, I would say 20$ per month max in api token costs.

Disclaimer: I am a Linux infra/k8s guy, I write production code but it's mainly glue code and mainly in golang.

Addendum: most value we get is from "document intelligence" and that's all Gemma and Qwen on H100/H200

New comment by dada216 in "AI is slowing down"

dada216 — Tue, 09 Jun 2026 13:16:45 +0000

Am I missing something or does the author not count the existing revenue to cover the buildout? I think they have it covered even in the worst case scenario (and requiring every piece of cash flow they have).

Also, if Anthropic or OpenAI fail in their revenue requirements I believe they would be absorbed by the guys with the money printing machines (Google, Microsoft and Meta).

New comment by dada216 in "NPM packages from RedHat have been compromised"

dada216 — Mon, 01 Jun 2026 14:12:41 +0000

not really, no.

New comment by dada216 in "Codex just found a "workaround" of not having sudo on my PC"

dada216 — Mon, 01 Jun 2026 08:51:06 +0000

I would also add buildah and skopeo to the mix of things that podman does better. also, podman desktop has better licensing that docker desktop. podman is modular and as such they could easily change the way they do networking over time, for once it doesn't break iptables and firewall rules by design but rather works together with the security design around these tools.

New comment by dada216 in "Codex just found a "workaround" of not having sudo on my PC"

dada216 — Mon, 01 Jun 2026 08:45:26 +0000

wait till you learn what the docker socket and the API can do and how container can get access to it. docker made bad design choices from the very beginning, mainly the API was designed coarse grained and they use a socket to do things, which is also easily exposed through the network you should not run docker on servers but rather use a better designed container runtime, docker rootless is a thing but it's been bolted on the bad design choice. podman, cri-o, containerd are all better options for servers hosting prod containers. it's kinda fine when you are using it on your dev machine, but I still rather take podman because I find it cleaner, buildah and skopeo are pretty useful as well.

New comment by dada216 in "Codex just found a "workaround" of not having sudo on my PC"

dada216 — Mon, 01 Jun 2026 08:35:36 +0000

which is what podman does with pasta

New comment by dada216 in "If you let AI do your writing, I will come to your house and kill you"

dada216 — Mon, 25 May 2026 10:24:06 +0000

he rants about the real world being ruined by ai as well, as opposed to just the digital one, then he goes on about needing catering services and, crucially, searching on google for them. my brother in bbq, why did you not go to a restaurant, a bbq joint, a food truck? that's the real world, searching on google is the digital world. bad rant, there's many critiques to be done to LLMs impact on the world, digital and real, but to me this guy does not seem mentally equipped to do them.

New comment by dada216 in "ASML staffing changes could result in a net reduction of around 1700 positions"

dada216 — Wed, 28 Jan 2026 10:38:17 +0000

It works well in finance, that's about it. It makes no sense, as you say, everywhere else where the core business has to do with anything that needs engineering.

New comment by dada216 in "Man shot and killed by federal agents in south Minneapolis this morning"

dada216 — Sat, 24 Jan 2026 19:52:11 +0000

quick reminder that out of the 8 author of the seminal paper that arguably started the whole LLM thing (attention is all you need) only one is American. all the other are foreign born, studied and worked abroad and only then were recruited by US FAANGs.

New comment by dada216 in "Gas Town's agent patterns, design bottlenecks, and vibecoding at scale"

dada216 — Sat, 24 Jan 2026 13:30:46 +0000

Embrace and use it to your advantage. Tell them nobody knows and understands how these things will actually work long term, that's why there's stuff like gas town, and that the way you see all of this is you can manage this process. What you bring to the table is making sure it will actually work if the tech is safe and sound, reaping the rewards, or protect the business if the tech fails, protecting the company from catastrophic tech failure, telll them that you are uniquely positioned to carry out the balancing act because you are deep in the tech itself. bonus if you explain the uncertainty framing in the business strategy: "because nobody really understands the tech nobody has an advantage, we are all playing on a leveled field, from the big boys at FAANGs to us peasants in normal non-tech enterprises: I am your advantage here if you give me the tools and leverage I need to make this work". if you play this right you'll get the fat bonus whether the tech actually works or not.

New comment by dada216 in "Microsoft CEO resorts to blogging in defense of AI"

dada216 — Wed, 07 Jan 2026 08:51:43 +0000

I actually thought it was a decent post. Corporate speech for "we may start to see some real world work being done with this, still gotta figure it out though"

New comment by dada216 in "Richard Stallman on ChatGPT"

dada216 — Tue, 09 Dec 2025 15:11:38 +0000

The question about what sort of "cognition" is going on in the simulated neural network of the current LLMs is an open one.

Saying "it's not intelligence" it's the wrong framing.

also, there are fully open LLM, including the training data.

He was right on a number of things, very important ones, but he's loosing it with old age, as we all will.

New comment by dada216 in "Go 1.25 Release Notes"

dada216 — Wed, 13 Aug 2025 09:11:09 +0000

anal_reactor started with "once I read on 4chan". that's all you should need to know to skip this entirely