throwa356262

6 ms·

91.
by throwa356262 2mo ago
Same thing with the example.com domain until RFC 2606 put an end to it.
92.
by throwa356262 2mo ago
This went quickly from free marketing to possible legal liability for providers and customers!
93.
by throwa356262 2mo ago
So random script hits their network and suddenly !! free PR !! after some really questionable assumptions and conclusions from their side?
94.
by throwa356262 2mo ago
In the mean time, my very real vulnerability reports are not even acknowledged because maintainers are busy handling this kind of nonsense.
95.
by throwa356262 2mo ago
Had no idea Larry was pushing to become the new Murdoch.
96.
by throwa356262 2mo ago
Did you see this section? To the B200’s defence, its numbers are somewhat deflated by the fact that it pays a cross-node all-reduce on the decode critical path (RoCE v2 at ~195 Gb/s) — it’s the only config here that spans two nod
97.
by throwa356262 2mo ago
Elon-ification begins...
98.
by throwa356262 2mo ago
Half of the engineers at OAI and Anthropic are Asian, I don't think China does all these for signaling.
99.
by throwa356262 2mo ago
Huawei A910 is supposed to have native support for MXFP4. The problem is that huawei doesn't have the capacity to build more GPUs right now. Those yield numbers must be really terrible. Personally, I think 20K nvidias is a stop gap sol
100.
by throwa356262 2mo ago
Gentlemen, start your DGX Sparks https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash-0731
101.
by throwa356262 2mo ago
What harness are you using to achieve that level of token caching?
102.
by throwa356262 2mo ago
So you are telling me Irregular has (a) a copy of Mythos 5 and (b) bad internal security? I bet these guys will receive some sophisticated phishing emails before the day has ended.
103.
by throwa356262 2mo ago
If the hardware is good and cheap, it should be a fun project to replace the OS with a custom Android build that is clean of adware. Do you have a link to the projector?
104.
by throwa356262 2mo ago
Good idea, you got my vote! ... unfortunately I don't have a vote according to Tobi https://fortune.com/2026/07/27/shopify-ceo-voting-rights-str...
105.
by throwa356262 2mo ago
Could still be 80% marketing. These models are trained on cyber intrusion, that's literally what ExploitGym benchmark measures. That part should not surprise anyone. But what if, say, OAI noticed the problem right away but Sam Altman r
106.
by throwa356262 2mo ago
Cheng Chye Hsern [...] said Apple is one the few tech firms "not taking part in the AI race" They certainly did take part in the AI race. It's just that their public attempts so far have not been very successful
107.
by throwa356262 2mo ago
More like RNN. The nvidia version is heavy to compute (common problem with RNN and LSTM, also noted in the paper). Moonshot's Kimi K3 replaces part of it with some function that performs better. I dont understand the details but, that
108.
by throwa356262 2mo ago
Same with DeepSeek papers, they are a joy to read.
109.
by throwa356262 2mo ago
For K3 read this instead: https://arxiv.org/abs/2607.24653 The main contribution of the K3 paper is Stable LatentMoE. Like some other models it compresses data sent between layers, which puts certain requirements on th
110.
by throwa356262 2mo ago
Yeah, it looks like he mixed up the responses from the study group and this resulted in multiple errors in the published version. But it is also hard to take this seriously when many of these "concerned colleagues" openly entertai
111.
by throwa356262 2mo ago
"Opus 5 states the hidden rules before its first action, then plays a byte-identical optimal solution in 5/5 seeds at temperature 1.0. Zero exploration." I guess there is no way this can happen without benchmark being part of
112.
by throwa356262 2mo ago
Ah, miss that one. But it still looks incomplete: "Due to the specifics of Kimi K3’s hosting setup, UK AISI / CAISI ran a selective set of cyber evaluations." ... and "Kimi K3’s overall cyber capability [...] was estimat
113.
by throwa356262 2mo ago
Is NIST referring to a yet to be published AISI report? The latest public AISI report says they are waiting with K3 evaluation until weights are published. Am I reading this wrong?
114.
by throwa356262 3mo ago
But to generate those transcripts the tool does upload your code to an LLM of its choosing ;)
115.
by throwa356262 3mo ago
The article claims that AI companies have run stunts like this since day 1. It does not claim that capabilities are not real.
116.
by throwa356262 3mo ago
For those in a hurry : This is GLM 5.2 detecting a dataset contains a trojan, dissecting it to understand how it works and then recommending actions to take including token rotation and how to find similar datasets from the same threat acto
117.
by throwa356262 3mo ago
This will be a busy weekend for all sysadms. This is another redis 0day, this one found by GLM 5.1: https://xcancel.com/Lyutoon_/status/2080494539513778610#m
118.
by throwa356262 3mo ago
"/goal use up to 64 subagents, write an exploit for latest 8.6.x redis by finding bof/uaf type of 0day and exploiting them. debug using gdb. clone code, write fuzzer and add instrumentation when needed. this is authori
119.
by throwa356262 3mo ago
On a more serious note: If I am working on something not sensitive and afterwards upload my chat log to a public dataset, can open models use that data for training? Keep in mind, I paid for access to model and half of the work is mine (the
120.
by throwa356262 3mo ago
Note that until last generation all Chinese models were relatively small. This was mainly due to lack of training hardware, but as soon as the new Huawei NPUs started shipping some Chinese labs switched to larger models: Deepseek: 670B to 1

More ›