HNHacker News
TopNewBestAskShowJobs

linzhangrun

373 karma · joined March 20, 2026

Developer in China. GitHub: https://github.com/Water-Run
submissionscomments
linzhangrun··on The unbearable cheapness of open weight models
Another thing these AI giants should worry about is the local deployment

A $4,699 DGX Spark can easily run Qwen 3.6 35B, and its performance crushes GPT-o1, the SOTA from two years ago

linzhangrun··on The gap between open weights LLMs and closed source LLMs
USA, a country that known for the land of freedom, is now restricting frontier models to the point where non-Americans cannot even use them.

China, a "authoritarian state" country, "the antonym of freedom", with a software industry that is especially capitalist, has produced all the competitive open-weight models.

It really is IRONIC.

Disclosure: I am Chinese, and I understand this strategy comes from being behind, using open source as an asymmetric way to compete and make up for missing compute by sharing the burden, etc. But still, very ironically.

linzhangrun··on How to corrupt an SQLite database file
Interesting title for official SQLite documentation :)
linzhangrun··on Why does kinetic energy increase quadratically, not linearly, with speed? (2011)
IIHS video shows the relationship between kinetic energy and speed in a very intuitive way:

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

For these basic virtual car experiments, BeamNG.drive is a pretty good physics simulator. You can open its built-in tools and run braking tests directly.

linzhangrun··on Previewing GPT‑5.6 Sol: a next-generation model
I saw videos of coding with Mimo-V2.5-Pro UltraSpeed, which is advertised at 1,000 tokens/s, which is very impressive.:

https://www.bilibili.com/video/BV1fME16uEW7

If the time-to-first-token latency also greatly improved, this could be very useful for end-to-end in controls, like autonomous driving for example.

linzhangrun··on Ask HN: Where is the programming profession going?
Your comment makes sense. Today's DGX Spark can easily run qwen-3.6-35B, which is indeed much stronger than the gpt-o1 two years ago. The future is hard to predict, perhaps the bigger issue is that while production capacity exists, where the demand will come from.
linzhangrun··on Show HN: I made Google Trends for Hacker News by indexing 18 years of comments
Great job! I've also been wanting to do similar statistics recently, wanting to know when LLMs becoming the absolute dominant topic on HN. Now it seems like half of the posts were about LLMs.
linzhangrun··on Google's Gemini 3.5 Pro Release Slips to July
Gemini really needs 3.5 Pro to prove itself. Feeling pessimistic about Gemini's future.
linzhangrun··on Half-Life 2 in a Browser
lmao :)
linzhangrun··on Ask HN: Where is the programming profession going?
Using today's model prices as a rebuttal is a very weak argument.

Two years ago, SOTA was gpt-o1, and it was much more expensive than Fable. Now, for $4,699, you can easily run a much smarter Qwen3.6-35B locally with DGX Spark.

Think about where we are. This is an era where a new SOTA arrives every two months. It took LLMs only about 18 months to go from chain-of-thought reasoning to disproving the unit-distance conjecture. chatGPT itself is only three and a half years old.

DeepSeek V4, released two months ago, is almost as cheap as the electricity costed, has the ability to being absolutely a top-tier model in 2025 standards.

linzhangrun··on Ask HN: Where is the programming profession going?
"Computer" use to be a job title. So no, I am not optimistic about the future of most programmers, maybe even all programmers.

One possibility is that software starts to look more like traditional manufacturing.

The machine is the company’s core asset. The engineer only needs to know how to operate the machine well. Once that happens, the barrier gets much lower, need much less people, and the job naturally become much less valuable. Some parts will still need to be done by hand, of course. But only a very small part. It is like old factories. They used to need lots of fitters, at all levels of skill. Now you only need a few of the elite ones.

AI is the CNC machine of the software industry.

The more pessimistic future is that, maybe five years from now, the best programmers will look at AI the same way the best Go or chess players look at AI today: Like KeJie said, "I don't even know what I am trying so hard against." We now have a new SOTA every two months. It just took 18 months for LLMs from reasoning models to disproving the unit distance conjecture. ChatGPT itself has not even existed for as long as a college student spends in university.

In any case, we have already passed the point where this can be rolled back.

Maybe ten years from now I will be leaving a comment saying that "programmer" used to be a job too :-/

Programming is the low-hanging fruit for AI. Open source and knowledge sharing have given it huge amount of public, high-quality training data at a level other industries can hardly imagine. And almost everything in programming can be tested and verified inside the computer quickly in a closed loop. No robot arm is needed.

The main weakness of current LLMs is still that they are static: They do not really change themselves through use. Harness tools are just elaborate ornamentation on top of prompts. LLMs are frozen at the moment training stops. Once we get models that can change their own weights through self-feedback, then maybe AGI really is on the horizon.

Thinking optimistically: I may be lucky enough to see it in my lifetime. Maybe by then, people will be able to live more like human beings, instead of organizing their whole lives around work :-)

linzhangrun··on There is minimal downside to switching to open models
And GLM-5 is built on DeepSeek V3 :)

There is still a lot of optimization left to squeeze out of LLMs.

linzhangrun··on What I'm Finding About LLM Code Style and Token Costs
why still use gpt-4o?
linzhangrun··on The unbearable cheapness of open weight models
It would not be surprising if GPT and Claude get cheaper too as inference gets cheaper. Two years ago, o1 was the strongest model and cost much more than Fable, while being nowhere near as smart as a Qwen 3.6 35B that you can now run on a DGX Spark without much trouble.
linzhangrun··on GLM-5.2 is a step change for open agents
Just waiting for the 3.5 Pro they said would come out this month. Gemini is pretty much useless for any serious work right now.
linzhangrun··on LuaJIT 3.0 proposed syntax extensions
Lua pursues "simplicity, purity, and simplicity." So... too much syntactic sugar is unlikely
linzhangrun··on GitHub shouldn't be a dependency for publishing Rust on crates.io
Cargo tied itself to GitHub back when GitHub still looked like an open-source utopia. Now the dependency is deeply baked in, and rolling it back would be very hard.
linzhangrun··on GTA 6 will cost $80
For GTA 6, I think $80 is acceptable. I just spent almost $100 in DCS mainly for the F-14B(U)(though that comparison is not really fair, since DCS is much more niche and naturally needs a higher price to make up for lower sales)

I also don’t think the extra $20 will hurt GTA’s sales much. What I’m more worried about is whether other 3A games will use Rockstar’s price increase as an excuse to raise their own prices . And I’m even more worried that GTA may come out with a pile of DEI slop, which would really feel like another Atari crash.

linzhangrun··on GTA 6 will cost $80
I don’t think Rockstar has ever launched on PC day one
linzhangrun··on LuaJIT 3.0 proposed syntax extensions
I thought luajit had completely stopped feature updates
linzhangrun··on Show HN: Y – A malleable coding-agent desktop app built with Electron
smalltalk is basically the textbook ancestor of OO
linzhangrun··on The deadly rise of giant trucks and SUVs
Look at those BEVs in the Chinese market — they're truly scary, 3-ton heavyweights running all over the streets.
linzhangrun··on Will It Mythos?
Google said they would bring 3.5 Pro this month. I've been waiting for a month now.
linzhangrun··on Jobs and Software Is Fucked
That kind of work needs robots in the loop. There is very little training data, most of it is private or opaque, and a lot of the know-how was never written down. It should hold out at least 20 years longer than programming, where basically the whole job happens inside the computer, and where the best references, examples, and source code are public to a degree unimaginable in most other industries.
linzhangrun··on Memory crisis is getting so bad that even retro RAM prices are going to the Moon
DDR3 era processors are severely outdated today, with the strongest d3cpu 5th gen i7 unable to match performance of 6W TDP N100
linzhangrun··on Runing GLM-5.2 on local hardware
Physical limitation of the manufacturing process may be more significant factor, starting from the TSMC 10nm ten years ago
linzhangrun··on Codex logging bug may write TBs to local SSDs
Considering the current storage prices and the SSDs whose lifespan you would thus exhaust...
linzhangrun··on GLM 5.2 vs. Opus
Out of stock, unavailable
linzhangrun··on GLM 5.2 vs. Opus
Just that their Coding Plan is too hard to get. I've been trying to grab it for a week and still can't get it
linzhangrun··on There is minimal downside to switching to open models
deepseek-v4-pro, probably the representative cheap opensouce LLM, was released in 2026.4 One year before, what OAI had in hand was gpt-4.1 and gpt-o3. I think it is not very controversial to say that deepseek is stronger than them, at most you can point to some post-training problems, basically the instability you mentioned. Also I am not sure if it is because the people who are best at using AI -- the people making AI -- get more development speed as the models get smarter, but my feeling is model progress is getting faster and faster. GPT-3.5 and GPT-4 were almost one year apart. The disadvantage from hardware limits and compute shortage is visible from the size of chinese models. glm-5.2, which is claimed to be around opus-4.6 level in coding, is only 744B. But Chinese engineers are obviously, how to put it, getting very effective results on "performance at the same size". And that is not even talking about the advantages from China's electricity, manpower, or even "national will" to compete against America. So saying it may take three years to catch up with a gap that is now only several months looks too pessimistic. ChatGPT itself was released only three and a half years ago, and today is already a completely different world.
← PreviousPage 5 of 6Next →