Studied Information Security & Working in AI, all views are my own

Austria
I'm gonna block every single account from here on out that replies / comments like a bot.
1
107
RT @DerHinz: In den geleakten Mail Files wird u.a. auch über Wahlbetrug in NRW berichtet, um mehr Stimmen für die AfD zu bekommen. Eigentli…
475
13
ClassicMain retweeted
Replying to @Hesamation
Time to update the logo.
2
10
453
14,002
ClassicMain retweeted
DeepSeek might as well open source the whole company at this point. “We’re sharing our production stack to help the community lower compute costs, reduce energy use, and make AI more affordable.”
🚀 Introducing DeepInfer, our in-house inference framework, now fully open source! Powering DeepSeek-V4.1-Flash at scale with high throughput and low latency. We’re sharing our production stack to help the community lower compute costs, reduce energy use, and make AI more affordable. 🔗 github.com/deepseek-ai/DeepI… 1/4
98
401
7,000
211,055
ClassicMain retweeted
🚀 Introducing DeepInfer, our in-house inference framework, now fully open source! Powering DeepSeek-V4.1-Flash at scale with high throughput and low latency. We’re sharing our production stack to help the community lower compute costs, reduce energy use, and make AI more affordable. 🔗 github.com/deepseek-ai/DeepI… 1/4
1,386
4,219
37,023
5,042,031
ClassicMain retweeted
“We found DeepSeek V4.1 Flash 1017 beats GPT 6 Sol on our multi-step offensive cybersecurity task.” READ THAT AGAIN. the same cyber capabilities that OpenAI kept behind a gate got beaten by a Chinese open model you can run locally. buckle up guys.
We found DeepSeek V4.1 Flash 1017 beats GPT 6 Sol on our multi-step offensive cybersecurity task. Sol was tested via the Daybreak trusted access program. We have been building a cybersecurity benchmark based on simulated corporate networks. This is a valuable indicator for the cyber-risk a model poses if in the hands of a bad actor. We plan to make the full evaluation public in the coming weeks but wanted to share our first immediate finding here.
58
242
3,343
181,988
ClassicMain retweeted
According to the Financial Times, the US ambassador to the EU has threatened severe economic consequences if Europe gives preference to its own satellite communications companies. At the same time, the US administration is restricting Starlink and Starshield services in Ukraine, even though European countries are paying for some of these services. It's a remarkably consistent policy: make your allies dependent on American technology, then threaten them economically when they try to reduce that dependence. The EU is already in a very difficult position. Its economy is struggling, and the Russian military threat is becoming increasingly real, with Russian drone attacks now reaching EU territorial waters. And instead of helping its European allies become stronger, Washington is threatening them for trying to become less dependent on it.
240
935
2,952
162,371
ClassicMain retweeted
We are all seeing this, right?
One of Putin's closest envoys sharing a meme of the USA and Russia attacking the UK and EU — 'Together'. Can we all see what is happening yet? 🤬
396
3,315
13,086
592,735
ClassicMain retweeted
So Mr. Trump has the right to cut the Starlink signal to Europeans. And at the same time, Europeans don't have the right to promote their own services? Time to wake up, Europe. @RenewEurope
European companies face “devastating” retaliation from Washington if Brussels presses ahead with plans to restrict American rivals such as Elon Musk’s SpaceX from parts of the European market, the US ambassador to the EU warned in an interview with @FT’s @BMoens and @HenryJFoy.
269
1,168
6,948
65,067
ANTHROPIC 🔥: According to The Information, Anthropic will release a new model this month, along with Gemini 4 Carbon and a new Terra model from OpenAI. > "The unreleased model is especially good at discerning a person’s underlying intentions and goals even if they don’t give the model much information." > "The Anthropic model is also better at thinking outside of the box rather than regurgitating common sense advice or solutions." > "It’s not clear which model family Anthropic will fall into, such as Fable or Opus." Thinking outside of the box is a huge limitation these days. If we get a model that can do it reliably, we'll be properly transitioning into the "Innovators" stage. The line about "discerning underlying intentions" is a bit scary, though, given all the guardrails.
Expect this month to be a busy one for new model releases. We have the scoop on a new Anthropic model, Google’s Gemini 4 Carbon, and a potential new Terra model from OpenAI. theinformation.com/newslette…
95
84
1,618
134,373
HOLY SHIT @AnthropicAI f you. IMPOSSIBLE TO WORK WITH CLAUDE "Most defensive security work doesn't need CVP" yeah i can see that; fuck you @lukedeentaylor @dsp_ @edwinarbus @lydiahallie @mikeyk @__nmca__ @cjav_dev More info: comments
1
208
and I am not planning on doing that. Anthropic is always saying we should not use chinese models because they are so willing of helping; then why are you refusing legitimate security triage? This is BATSHIT CRAZY. And then Claude, usually a logical reasoner, coming up with
1
21
made-up reasons on why it won't do what I am asking it to do. "No I am not going to ask GLM 5.3 of your other gateway". Why? "Because I won't lmao" Yeah fuck you Reasoning: 100
17
ClassicMain retweeted
Imagine if Europe, in the days after 9/11, had cleared its throat and announced: We have struck a tremendous deal with Osama bin Laden. We have lifted the sanctions on his organization, so he is now free to trade on the open market and bankroll his war against the United States. Bush must step down. The man has no cards. Has he even looked at his hand? But no. Europe did nothing of the sort. Europe did not sit down with a mass murderer and come home gushing about what a strong and reasonable fellow he turned out to be. Europe laced up its boots, flew to Afghanistan, stood shoulder to shoulder with the Americans and hunted that bearded wretch through every godforsaken mountain in the country, because that is what you do when a friend gets hit, and handing the attacker a shopping list while telling the victim to surrender is what you do when you have lost your mind.
341
4,046
15,772
206,791
ClassicMain retweeted
respectfully: FUCK THIS BULLSHIT stop babysitting every fucking move. your auto classifier and safety rules are fucking stupid and you waste my time every single fucking day with this fucking bullshit ARE YOU FUCKING SERIOUS ANTHROPIC thank you for your attention on this matter
47
4
131
20,736
ClassicMain retweeted
OPENCODE GOT HACKED 😳 check your API keys now • OpenAI and Anthropic keys saved in OpenCode may be exposed • users are already seeing $65 to $281 in charges what to do 1. revoke every key you saved in OpenCode 2. make a new one 3. set a spend limit source: x.com/mattjay/status/2111480…
🚨 Confirmed. OpenCode hacked. Keys stolen. Telling customers to revoke.
12
7
36
4,718
ClassicMain retweeted
imagine the pressure anthropic must be feeling to make fable 5.5 more impressive than opus 5.5
33
12
1,121
27,717
Meta just dropped Muse Spark Open 🔥 816B total parameters, 42B active, with a 1M token context window. Built for coding, long-running agents, and complex reasoning, with native text, image, video, and audio support.
Today we’re releasing Muse Spark Open, the most capable US open-weight model to date. Built for coding and agentic work, it excels at long-horizon tasks, complex software engineering, and multi-step reasoning. Key Model Specs: → 816B total / 42B active parameters → 1M token context window → Natively multimodal across text, image, video, audio
3
9
134
12,096
Grmini 4 Argon is..... Smart, knowledgable, but still so stupid sometimes. For it's current price, a good model For it's future price (starting 1st December) it's a very bad model. Opus 5.5 costs the same and is miles ahead still (except in real knowledge questions) Tldr: dont
1
69