ۗۗۗۗۗۗۗۗۗۗۗۗۗۗۗۗcyberflâneur @fileverse all watched over by A̶̙̗̩͖̺͂̀̑̑Ļ̶̡͎̥̭͇̝͙̻̪̭͚͖̾̏̍͛͜͜͠Ģ̴̭̥͙͖̲͖̮̺̬̼͍̻̌̆́̊̓̄̈́̎Ồ̸̘̟̟̳̰͇̥̾̋̂̋̈́̂͐͌̕͝machines of loving grace
n̵̨̧̧͙̗̗̗͉̳̪̥̺̳̠̠̖̪͉̭̥̼̹͙̝̣̗̜̣
- Tweets 9,638
- Following 1,014
- Followers 2,128
- Likes 82,856
In "today's" catastrophic security news:
Billions of Intel processors (i.e. most your devices and the Internet lol) have a critical weakness allowing attackers to steal passwords, encryption keys, private data, etc. Even SGX (specifically designed to protect peoples data) is affected
downfall.page
Libbitcoin Explorer (bx seed) –– used by many crypto wallets to generate private key entropy (even recommended in Mastering Bitcoin) is flawed and hackers have been silently draining people's money across multiple blockchains ¯\_(ツ)_/¯
milksad.info/
66
235
761
264,137
Pushing Bluetooth out to 12 Kilometers ift.tt/fNbw2BF
8
42
330
23,388
(⧖)E ه҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿ retweeted
Essa descoberta foi bizarra, mas vindo da NVIDIA que capava as GPUs na época de mineração de cripto, não é surpresa nenhuma.
Todas as 3090 estão com metade da capacidade bloqueadas por FIRMWARE. Ou seja, se você tiver uma e aplicar o patch, vai desbloquear a outra metade e ficar com uma GPU quase 2x mais potente que antes.
VSF NVIDIA.
Confirmed that 3090 unlock is successful
with 360watt power wall, I got 75T -> 135T (measured in Pytorch)
Which is insane
I used github.com/Zagitta/gpu-unloc… with commit a9b0c75
169
663
10,793
426,561
(⧖)E ه҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿ retweeted
Moonshot is taking a very interesting approach to cyber for its new Kimi 3.1 Pro model.
Instead of adding safety guardrails during post-training, they removed cyber-related data in mid- and post-training. This tries to address the problem of people fine-tuning away the safety guardrails for cyber for the open-weight model.
At the same time, Moonshot still has a more powerful internal version of the model with full cyber capabilities. Now, like the US, China’s true AI frontier is no longer public.
35
57
543
31,300
give me parametres, prompts, proofs. otherwise it just marketing.
written like that it could mean anything from "model copy pasted a known technique" to "model self-triggered and executed a full side-channel communication with other isolated agents"
completely insane
OpenAI agents communicated via cache timing differences, without any shared files or network connection
1
6
288
(⧖)E ه҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿ retweeted
Introducing Kimi K3.1 Pro: our most capable 3T-class open-weight model.
🔹 1M-token context, 2.8T total / 104B active params (MoE), Multimodal.
🔹 Full-project coding: games in Unreal, Unity & Godot.
🔹 Excels at parametric 3D modeling in Blender & CAD.
🔹 More reliable at finance work: Accurate modeling with hallucination down by 65%.
🔹 Matching the strongest proprietary models across coding, agentic and knowledge benchmarks.
Kimi K3.1 Pro and K3.1 Pro Highspeed (2–3× faster) are now live on Kimi.ai, Kimi Work, Kimi Code, and the Kimi API.
Open weights by Nov 4, 2026.
🔗 API: platform.kimi.ai
🔗 Tech blog: kimi.ai/blog/kimi-k3-1-pro
1,218
3,178
29,734
3,677,437
(⧖)E ه҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿ retweeted
"Google DeepMind researchers are under in perpetuity non-disparagement clauses forbidding them from ever speaking critically of the company."
Sorry wait what?
Important, and I'm glad this is finally public. Google DeepMind researchers are under in perpetuity non-disparagement clauses forbidding them from ever speaking critically of the company. I'd already confirmed this with multiple people.
Yes, some ex-employees have spoken out nonetheless. But it has also discouraged others from doing so. And yes, it most certainly has a chilling effect with regards to discussions on safety and risk.
Others have written compellingly about how important the ability to speak freely and critically is, especially on topics like safety and risk. And what a testament to the research employee community's coordination and hard-won bargaining power it has been to get there at other companies.
GDM burn that commons by having this. And everyone who signs burns that commons too. Let's get it changed.
6
26
463
19,135
(⧖)E ه҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿ retweeted
Anyone wanna take a guess at what’s going on here or has there now been enough indisputable proof that openai is monitoring the logs of high profile experts.
44
348
4,000
100,524
(⧖)E ه҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿ retweeted
The order banning DeepSeek in the US government is absurd. Why ban US-hosted DeepSeek? "The models themselves may contain malware".
The proof? The DeepSeek iOS app has vulnerabilities.
108
49
800
56,876
(⧖)E ه҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿ retweeted
I think people will generally underestimate how much technological acceleration is directly attributable to deepseek.
Because of them, it is going to proliferate. Really, it's because of them. Thank you, for everything
109
395
6,724
217,770
(⧖)E ه҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿ retweeted
OpenAI, Anthropic and other leading AI labs face a growing challenge: Better models aren't necessarily attracting more usage. A 10-point intelligence advantage now earns just 20% more developer traffic, down from nearly 69% in late 2025.
Full story: thein.fo/4hGJ9jY
47
82
1,068
558,257
AI will not solve data portability, it will make things worse.
Get yourself on open source apps NOW or you'll be locked in forever.
Given the acceleration Anthropic is reporting in coding, it is amazing that you still cannot change your email address in Claude and the official advice is that if you need to change your email the only way to do that is delete your account.
2
13
243
(⧖)E ه҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿ retweeted
This is actually more a side-effect of the death of pretraining.
As transformers saturate, most pretraining results are tweaking data recipes, not massive stepwise changes in performance.
Turns focus to innovation in postraining and test time compute. This is good for competition, since pretraining concentrated resources in a few labs.
GLM 5.3’s biggest leap came from post-training alone.
“GLM 5.3 was built entirely by post-training the previous model. For me, that was a wake-up call.”
“Post-training is far more GPU-efficient than pre-training from scratch. Plenty of companies can’t afford to train a model from scratch, but they have proprietary data. They can post-train themselves and beat some of the closed models.”
“There’s still so much juice left to squeeze from post-training. We compared GLM 5.3’s cyber capabilities with 5.2, and it’s a night and day difference. That’s all just post-training.”
28
40
378
39,817
Fileverse
*everyone panicking about AI breaking classical encryption*
claude: hey turns out the generation of ECDSA curve seeds is reproducible! we can worry a bit less about potential weaknesses
2
2
25
4,547
(⧖)E ه҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿ retweeted
A folder 𓅹
⠀└ inside another folder
└ inside ur ddocs.new
🍓 Sub-folders + document ⧉ duplication, now live.
9
17
98
3,616
(⧖)E ه҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿҈̿ retweeted
Thread 🧵
1/ NEW REPORT: Billions of advertising auctions determine in real time which ads we see on websites and in apps every day. Our latest research shows how the same data fueling this advertising ecosystem is being used to profile millions of users and silently infect targets with zero-click mercenary spyware, a novel attack technique we call "Adfection.”
Read the report: citizenlab.ca/research/how-d…
18
351
553
81,104