HackerTrans
TopNewTrendsCommentsPastAskShowJobs

bertili

no profile record

Submissions

Ask HN: Dunkelflaute' turns off my monitor

4 points·by bertili·il y a 7 mois·1 comments

comments

bertili
·il y a 24 jours·discuss
This is GLM 5.2 Max. GLM 5.2 High which use less than half[1] the tokens.

[1] https://z.ai/blog/glm-5.2
bertili
·il y a 28 jours·discuss
Qwen 27b is a compute heavy dense model.
bertili
·il y a 2 mois·discuss
Does this translate into a similar reduction in compute?

What's the catch?
bertili
·il y a 2 mois·discuss
equals 2 or 3 human brains in power usage. Amazing work!
bertili
·il y a 3 mois·discuss
It's fascinating that a $999 Mac Mini (M4 32GB) with almost similar wattage as a human brain gets us this far.
bertili
·il y a 3 mois·discuss
Is there any source for these claims?
bertili
·il y a 3 mois·discuss
A relief to see the Qwen team still publishing open weights, after the kneecapping [1] and departures of Junyang Lin and others [2]!

[1] https://news.ycombinator.com/item?id=47246746 [2] https://news.ycombinator.com/item?id=47249343
bertili
·il y a 3 mois·discuss
The timing is interesting as Apple supposedly will distill google models in the upcoming Siri update [1]. So maybe Gemma is a lower bound on what we can expect baked into iPhones.

[1] https://news.ycombinator.com/item?id=47520438
bertili
·il y a 3 mois·discuss
Qwen: Hold my beer

https://news.ycombinator.com/item?id=47615002
bertili
·il y a 4 mois·discuss
Very impressive! I wonder if there is a similar path for Linux using system memory instead of SSD? Hell, maybe even a case for the return of some kind of ROMs of weights?
bertili
·il y a 5 mois·discuss
Better than frontier pelicans as of 2025
bertili
·il y a 5 mois·discuss
Most certainly not, but the Unsloth MLX fits 256GB.
bertili
·il y a 5 mois·discuss
Last Chinese new year we would not have predicted a Sonnet 4.5 level model that runs local and fast on a 2026 M5 Max MacBook Pro, but it's now a real possibility.
bertili
·il y a 5 mois·discuss
Exactly. The emperor has no clothes. The largest investments in US tech in history and yet there less than a year of moat. OpenAI or Anthropic will not be able to compete with Chinese server farms and so the US strategy is misplaced investments that will come home to roast.

And we will have Deepseek 4 in a few days...
bertili
·il y a 5 mois·discuss
Surely this is the elephant in the room, but the point here is that Apple as control over its ecosystem, so it may be able to sandbox and make entitlements and transparency good enough, in the apps that the bot can access.
bertili
·il y a 5 mois·discuss
People are running the previous Kimi K2 on 2 Mac Studios at 21tokens/s or 4 Macs at 30tokens/s. Its still premature, but not a completely crazy proposition for the near future, giving the rate of progress.
bertili
·il y a 5 mois·discuss
The other realistic setup is $20k, for a small company that needs a private AI for coding or other internal agentic use with two Mac Studios connected over thunderbolt 5 RMDA.
bertili
·il y a 5 mois·discuss
The "Deepseek moment" is just one year ago today!

Coincidence or not, let's just marvel for a second over this amount of magic/technology that's being given away for free... and how liberating and different this is than OpenAI and others that were closed to "protect us all".
bertili
·il y a 6 mois·discuss
What a time to be (still) alive.

When we look back in a few years and ask the question: who actually got to pay for the Epstein crimes and coverups, we come to the surprising answer it is the Greenlandes and other innocent societies that got ripped apart by this maniac and his supporters.
bertili
·il y a 6 mois·discuss
It's not a failure of IP6 but a failure of society.

We all thought the internet would become decentralized and that everyone should have an IP and a funky website. But instead social media took over, big tech and a few big discussion sites where we all must fit in a digital life and watch ads and share our data to become a good product for all the others to consume.