Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
gunalx
searching Neon…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
6 ms
·
1.
▲
by
gunalx
3d ago
Bummer to not see the atx control support. Its the main use i have from my ipkvm to do fde decryption and boot the machine once wake on lan fails.
2.
▲
by
gunalx
4d ago
There is too many ads to report all of them. But a don't show me this, that also affected on report as a negative signal to reduce the possibility of that specific ad to you.
3.
▲
by
gunalx
6d ago
Can't wait till a customer rep is impossible to get to because all support is outsourced to ai.
4.
▲
by
gunalx
6d ago
Right. I got the feeling of it being saturated because all the top 5 fully completed it.
5.
▲
by
gunalx
6d ago
More often than not just knowing the contact has OK connection is good enough. As far as I know the marker chips is mostly to tell the devices to not abuse the cable.
6.
▲
by
gunalx
6d ago
Wouldn't work. A dynamic but verifiable problem. Is just a perfect target for a RL environment. If you don't have the verifiable part the benchmark is useless, or really expensive with human review. (Or just open ended)
7.
▲
by
gunalx
6d ago
In fact we should though.
8.
▲
by
gunalx
6d ago
Ignoring the obvious ai slop webpage. I don't really trust the benchmark. It seems either pretty saturated, or inconsistent just based on the results.
9.
▲
by
gunalx
7d ago
A casual look around the office. Not a good representative selection for the entire population but counting to 10 and asking, everyone used ai in some fashion, extrapolate the data and you get 100%. (Not a ai first office even)
10.
▲
by
gunalx
8d ago
Fact is most people don't really care. And they don't really know why they should care either. All the problems feels like something far away, while they are being exploited and sold without fully realising. I dont think anyone kn
11.
▲
by
gunalx
12d ago
I mean. Fundamentally they are just autoregressive next token predictors. Fundamently fou can simplify to f(x) -> x+y where x is input tokens, y is the next token and f() is the model function. Yes the model function is complex but still
12.
▲
by
gunalx
13d ago
The problem I see with vue is what vue patterns are you using. And the llms may be trained on more outdated ones more than current practices.
13.
▲
by
gunalx
16d ago
God that had a lot of llm'isms. Couldnt get through it with all the repeats and info bloat.
14.
▲
by
gunalx
20d ago
This is so funny. Would love a more readable font option in the same style though.
15.
▲
by
gunalx
21d ago
Just as expensive unobtanium as nvidia. Maybe even more so. Its probably the huawei ascend 910, they are using.
16.
▲
by
gunalx
21d ago
They are high end really expensive Huawei ascend GPUs. It is kinda bruteforcing the performance on a older semiconductor processing tech, so total production is pretty low.
17.
▲
by
gunalx
21d ago
Early glm models gave off that wibe. But now its more inspired by. With a bit of Claude in there. But I do think they actually do RL otherwise glm5.3 shouldn't have been able to beat fable on the few tests it did.
18.
▲
by
gunalx
21d ago
Yes, the terms are dubious. But they are also reasonably lenient with enforcement. They also don't require persona id verification, witch is wat turned me away from openai.
19.
▲
by
gunalx
21d ago
I have had the opposite where I felt 5.3 as a stronger model than 5.2. Its feels way more in tune with my code, and does more nuanced edits. Though I do handhold my models a lot, so might fall outside the agentic term.
20.
▲
by
gunalx
24d ago
glm 5.3 gets awfully slow during peak hours. But you might not hit them to frequently.
21.
▲
by
gunalx
27d ago
Cute. But I don't see the real uppsides here vs any other minipc.
22.
▲
by
gunalx
27d ago
Yes true. But also way to often neglected.
23.
▲
by
gunalx
27d ago
Single page applications are one of the most anoying web patterns i know of. Please just let me have one page for each ting to do so i actually can bookmark it properly.
24.
▲
by
gunalx
28d ago
Well that would not work as the metadata is also part of the sha, and if you have a sha predictably contain or hash to itself you have a bad hashing algorithm. (Unless of course you mean sha of the weigths and not the gguf itself)
25.
▲
by
gunalx
28d ago
Also almost any native desktop app forced to mobile has broken ux.
26.
▲
by
gunalx
28d ago
Yep. Had to change banks because of my old one suddenly dropping grapheme support by adding stupid attestation mechanics. They surely must have gotten feedback from me and others because the next update it worked again. But I and probably o
27.
▲
by
gunalx
1mo ago
So ill strangle myself while the 3d printer goes back and forth.
28.
▲
by
gunalx
1mo ago
Common on non us keebs though.
29.
▲
by
gunalx
1mo ago
+1 gemini models where really the only ones fullt grasping spatial reasoning even compared to opus (at least when i last cared to check it)
30.
▲
by
gunalx
1mo ago
Personally both, i use openscad with opencode, and paste images but it is pretty often the modell decides by itself it wants to see a render and uses the render shell commands to get a image to look at.
More ›