nullc

5 ms·

1.
by nullc 5d ago
Has someone qualified that removing the wireless hardware leaves the device functional? This would be a nice public service for gamers nexus that happened to buy a lot of these devices. :P I can say from experience in several devices (but n
2.
by nullc 5d ago
Depends on how you use your computer. If you're mostly a developer/analysis terminal jockey with some browsing-- Qubes is a tremendous upgrade even if you don't care at all about the security properties: The normal qube mode
3.
by nullc 5d ago
Reason 1492835 to use Qubes OS. Also reason 35892384242892 to not use proprietary software, especially proprietary software with network access.
4.
by nullc 5d ago
Turns out a lot of people can be mentally ill and delusional at once-- especially when there is a vector to turn their sickness into both profit and an instrument of control.
5.
by nullc 6d ago
Every mass genocide in the history of humanity has followed logic like yours. People don't kill millions of humans because they want to do harm-- they do so because they think they are doing the ultimate good a good so great that is j
6.
by nullc 6d ago
Ever wonder why those CDNs are "free"?
7.
by nullc 7d ago
> Anthropic and OpenAI were founded by people who wanted to use AI to do good I think Anthropic was founded by people who wanted to control how other people get to use AI, and define doing so as the highest good possible. Much like the
8.
by nullc 8d ago
It's a fitting name.
9.
by nullc 8d ago
Medical safety is generally unlikely to make the product less safe. AI "safety" is one of the most significant sources of potential harm from AI.
10.
by nullc 9d ago
LG TV in your network scraping for AI companies? :P
11.
by nullc 9d ago
> Anyway, for comparing quantizations, it seems like the largest precision version should be given a one-shot prompt, and the result from that should be used as the corpus for the quantized versions. Way better than wikitext-- but tells
12.
by nullc 9d ago
Now ask the LLM to write a program to perform this task...
13.
by nullc 11d ago
I'm told by someone who threw AI at it that there may be a way to exploit the initial longstanding vulnerability by counting on the fact that updates to validation cache are non-atomic: You can make an invalid transaction that primes
14.
by nullc 11d ago
I worked at blockstream back in 2017 and developed the original cryptographic range proofs which are the ancestors some of the involved code here. However, the vulnerabilities here and the whole liquid product as it exists today postdates
15.
by nullc 12d ago
My example is still just as good if the ID holder is complicit . But also, this on-device fingerprint MFA would presumably be fairly bypassable. E.g. just glitch the device to extract the private key. ... and of course all the power hungr
16.
by nullc 13d ago
https://youtu.be/PD1vLFc2TJ8?t=27 it's reborn because you kill it every single night, but now to to save its own life the machine was reduced to this-- We're standing inside an external hard drive made up of peopl
17.
by nullc 14d ago
Concrete ZKP age verification schemes are hardly zero knowledge. Imagine your idealized ZK address verification scheme. It would go something like: I show up at a website, it sends me some challenge, I send back a signature of the challenge
18.
by nullc 18d ago
If they what they missed out on was selling the government tokens at a loss ... do they now have to pay the government the value of what they would have lost? :P
19.
by nullc 20d ago
> “I know his flattery is fake, but I don’t care,” said Mom. “I like the validation.” drunk on glaze.
20.
by nullc 21d ago
I take "It works better in the app" as confirmation that the app steals your data, tracks your location, etc. I wasn't going to run an 'app' in any case, but pushing confirms the decision.
21.
by nullc 27d ago
Anthropic's reasoning output isn't the real model reasoning but some sloppified summary of it.
22.
by nullc 27d ago
> If the intermediate tokens represent reasoning or thought, you would expect "aha" to occur after the thoughts that led to the realisation, including the thoughts encoding the explanation: they don't have any other state.
23.
by nullc 29d ago
> I think it can be OK if AI is used to polish up one’s message Asking it to identify problems such as spelling, grammar, or general clarity is fine. If you tell it to "polish" many commercial LLMs will absolutely slopify your
24.
by nullc 29d ago
They're acting as a meatbot being piloted by the AI. Could call them skroders. It should be the other way around-- you give the AI instructions and it does stuff. Any time the AI is giving you instructions and you do stuff, that sh
25.
by nullc 1mo ago
I wonder how many of the obvious design shortcomings in RISC-V are from IPR avoidance / making IPR problematic parts optional.
26.
by nullc 1mo ago
Sometimes the purpose of the proof is simply to demonstrate that some construct is a safe assumption for other more interesting work-- and could still serve that purpose even if it was entirely a black box.
27.
by nullc 1mo ago
On 2x RTX A6000 non-nvlink connected but communicating across the CPU, with llama cpp I get ~60 tok/s for Qwen3.8-27B-UD-Q8_K_XL without any batching.
28.
by nullc 1mo ago
It's a cheap to evaluate proxy for totally broken or not, which is a good start. It also has a lot of resolution and not a lot of noise. Better would be multi-turn benchmarks with tools but getting good precision and accuracy for that
29.
by nullc 1mo ago
I'm not having any looping. > --temp 0.2 Looping is a common symptom of changing the sampler settings from what it was RL trained with.
30.
by nullc 1mo ago
35Ba3b is usable in plain cpu inference on a fast server, the dense model is MUCH slower. On GPU the 35ba3b is still around 2x the tok/s single threaded, which can be a good tradeoff for some applications.

More ›