bertili

7 ms·

1.
by bertili 7d ago
The bigger story is the compute efficiency - its been running at 300t/s the last days.
2.
by bertili 14d ago
DeepSWE scores 75.4 - that's the best score so far. And it's crazy cheap! Google held the top a few hours today with Gemini 3.8 Flash, but now second to Spark 1.3. All this competition will drive prices down!
3.
by bertili 14d ago
A fifth of the cost of Opus 5! Google is certainly pushing the completion with this.
4.
by bertili 21d ago
0 days later: Qwen 3.8 Flash Next: Let's cut GLM 5.3 Flash parmeters in half and active parameters to a third! Chinese models had 94% reduction in parameters (from 2.8T/104B to 180B/6B) in 6 weeks, while staying close to the
5.
by bertili 21d ago
The point is open AI. "Open" as in open weights, open research, open future.
6.
by bertili 21d ago
This is going so fast! What a time to be on hackernews: July 16th: The "Kimi K3 moment" - China has caught up to Opus! 4 weeks later: GLM 5.3 - Same performance, but cut the amount of parameters and cost to a third! 12 days later:
7.
by bertili 1mo ago
I can't shake this the existential feeling that this compact series of 27G bytes represent something profound and universal.
8.
by bertili 1mo ago
And more context: Same score as the latest DeepSeek Flash 0731 which has 284B parameters! (13B active) Its also the second best Qwen model, much better than Qwen 3.7 Max, but significantly below Qwen 3.8 Max.
9.
by bertili 1mo ago
Wow. Speed improved as well. 200t/s on a RTX 5090! https://x.com/sgl_project/status/2088281320422322413
10.
by bertili 1mo ago
This will be roughly on pair with Kimi K3, but using a third of its parameters. Just 4 weeks ago the "Kimi K3 moment" was seen as a threat to Closed AI and in less than a month Z.ai have cut the parameter/RAM barrier to a thi
11.
by bertili 1mo ago
Musk: Open Chinese models will rival Fable 5 in Q1 2027 JieTang (Founder of Z.ai): It won't take that long https://x.com/i/trending/2067626647050670400?lang=en
12.
by bertili 1mo ago
DwarfStar ( https://github.com/antirez/ds4 ) supports GLM 5.2 and DeepSeek. Not only for toying, but for getting work done.
13.
by bertili 2mo ago
The 27B have many more active parameters than much bigger models such as DS4Flash, MiniMax etc, which makes it punch above its tiny weight. A great fit for a 5090 in a closet for meat-and-potatoes, kind of work.
14.
by bertili 2mo ago
That looks promising! As models become a commodity, this may turn out to be the real AI gold rush.
15.
by bertili 2mo ago
Is there any (near future) technology that would permit burning this terrabyte into some kind of ROM chip?
16.
by bertili 2mo ago
Wait.. the Qwen Max models have never been open-weight. But it sure sound like that's what they intend now? "Qwen3.8 is launching and going open-weight soon! With a massive 2.4T parameters..."
17.
by bertili 2mo ago
AGI is almost here, but first, one more thing... a keyboard controller!
18.
by bertili 2mo ago
Legislators, please require 10 seconds of load screen with a picture of a tree, for every online video. It worked for cigarette packs.
19.
by bertili 3mo ago
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
20.
by bertili 3mo ago
Qwen 27b is a compute heavy dense model.
21.
by bertili 4mo ago
Does this translate into a similar reduction in compute? What's the catch?
22.
by bertili 4mo ago
equals 2 or 3 human brains in power usage. Amazing work!
23.
by bertili 5mo ago
It's fascinating that a $999 Mac Mini (M4 32GB) with almost similar wattage as a human brain gets us this far.
24.
by bertili 5mo ago
Is there any source for these claims?
25.
by bertili 5mo ago
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.
26.
by bertili 6mo ago
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=475
27.
by bertili 6mo ago
Qwen: Hold my beer https://news.ycombinator.com/item?id=47615002
28.
by bertili 6mo ago
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?
29.
by bertili 7mo ago
Better than frontier pelicans as of 2025
30.
by bertili 7mo ago
Most certainly not, but the Unsloth MLX fits 256GB.

More ›