Hacker Newsnew | past | comments | ask | show | jobs | submit | ThouYS's commentslogin

wait, the thing that streams my code into the cloud, and that I let run basically arbitrary commands on my machine... uploads my code into the cloud?! I didn't sign up for this!

yikes.png

if this beats GLM 5.3 flash, I am sold

i think this will beat GLM 5.3 flash, i mean the last ds4-flash after preview was already great, and i found that more intelligent than GLM 5.3 flash

Wasn't that part of their contract with Microsoft? Some clause stopped biting with the arrival of AGI


qwen is a magical model. it has the mandate of heaven. was so already at 3.6-27B


GLM 5.3-flash fits the bill


and $0.15 1M input, $0.50 1M output

I am a huge enthusiast of running local models, but when multiple quality USA vendors provide models like GLM 5.3-flash, I run locally just for the fun of it.

For the purposes of comparing to Fable 5.1, I would mention GLM 5.3 that is about 1/12 the cost.


What is it equivalent to?

What kind of things are you using it for?

I haven't tested it yet but on all the benchmarks it looks like it's 5-7x slower for agentic tasks.


I made some webapps with it, and have it running my hermes agent (which also does a lot of coding, but not webapps).

Not sure what it's equivalent to, but it's super cheap and I am happy with the results


It's a mix of slightly worse kimi k3 for UI work and slightly smarter than luna for everything else.

But yeah, it's very slow. I've put it to work as an LLM-as-RAG agent.


I was wondering that, when DeepSeek became so cheap a while back, if it would be suitable as a superior embedding model.

Although, RAG means search and search means latency?


This might be the ideal form factor for local+ models. Even though I am a diehard qwen3.6-27B fan, it is not the same as this guy. GLM 5.3 you can actually run on-prem reasonably well, and it is a legit, proper, work horse. As in, this can do real work.


ok this is nice


copyright etc has been dead the moment the personal computer arrived. we are still living through the chaos of the new world being born, but I think it has been over for a while already.

I would be very surprised if any kind of license like the GPL (or any license for published code) is left in 100 years or so.

DRM is another face of the same struggle, and this too shall pass. Ironically, DRM coming from the same source as the GPL (wanting to force rules onto something freely copyable), is viewed very harshly, since it supposedly interferes with your "freedom". But in the end they are the same.

Both GPL and DRM come from the idealistic fantasy that non-scarce digital goods have any (market) value at all. AI is just the latest nail in this coffin.


That's the whole point and philosophy behind the GPL. That's why it's called "copyleft". It's about taking copyright law as it currently exists and using it subversively.


That subversion is key to GPL but it's not the point. The point is to get the code shared as widely as possible.

The person you're talking to is missing that point because they're starting at "freely copyable", while the actual default is that code is never published at all. GPL is compromise to push source code 90% of the way to "freely copyable" because the alternative is usually 0% copyable.


Thanks Peter! Now do the same for the 'gram!


Why don't you do it?


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: