💻 IT-Admin & Automation Enthusiast Linux, Selfhosting, Open Source, Server und smarte Workflows
Österreich
Joined January 2019
- Tweets13.9K
- Following143
- Followers542
- Likes22.2K
P-09 retweeted
🚨GPT-6 Bel Coming Soon
>It's said that Bel is Doug's successor
>Reportedly, total parameters exceed 10T
>Bel may serve as the foundation for the model after GPT-6
>OpenAl reportedly views Bel as a potential post-GPT-6 foundation
>It might even become the foundation for an AGI threshold model
P-09 retweeted
Ubuntu 26.10 supports new authentication flows for Microsoft identity services, along with 26.04 LTS and 24.04 LTS. 🔒
Learn more: canonical.com/blog/canonical…
P-09 retweeted
I asked Jensen Huang how I should raise my children in the age of AI.
What do you teach your kids when you can’t predict the future they’ll grow up in?
His answer touched on curiosity, hard work, and being willing to struggle with something difficult.
A personal moment from our conversation. Full interview coming soon on Superintelligence.
P-09 retweeted
Still super confused how its possible that Grokbot got approval for EU but OpenAIs dot, Anthropics duet and Metas Muse not.
I mean, I take it. Grok bot is just insanely good. My daily always on agent(s) for almost every task. Its just so good and it gets better almost every day. Kudos spaceXSI.
P-09 retweeted
"Bill Gates says to prepare for 20 years of ‘tumultuous’ AI disruption".
I'm always annoyed by headlines like these. They suggest it's a foregone conclusion, as if chaos is inevitably about to descend upon us.
But the future is an opportunity; I see far more positives, more optimism than chaos. And how things unfold is in our hands, not a given.
P-09 retweeted
Good to know: OpenAI resumed RL training their frontier models on October 1st. It is necessary for OpenAI to keep up with Anthropic.
Over the past several months, we have substantially hardened the security of our research training environment. We’ve removed live internet access for all frontier models during training and built an offline world for them to use instead—a cached version of the web where they can still learn to browse, search, and use applications.
On September 20, after a model exploited a DNS misconfiguration to find a way outside of this offline world, we paused frontier training to strengthen containment further. We assembled teams of agents, powered by our most cyber-capable models, and tasked them with finding weaknesses in our research environments.
While they found no route to the internet they could easily exploit while concealing their activity from our monitors, they did find zero-day vulnerabilities, virtual machine escapes, and hardware side-channels for communication between agents. We used their discoveries to further strengthen our defenses and contribute fixes to open-source software.
After strengthening our containment architecture, we resumed our frontier RL training and inference on October 1. We’ve reserved substantial compute for continuous red-teaming, and made high-compute red-teaming part of the safety cases we use to decide whether frontier training and evaluation can proceed securely.
openai.com/index/securing-ou…
P-09 retweeted
Leute können sich nicht vorstellen, wie wild 2027 mit Voice Agents wird.
Full-Duplex Speech-to-Speech Modell, schneller als Echtzeit und lokal auf einem MacBook Pro M4 Max.
Sprich Parameter im einstelligen bis niedrigen zweistelligen Milliarden-Bereich.
Trainiert komplett ohne Text und ohne Pretrained Basismodell laut Block.
Wie Bruce Willis in The Sixth Sense: Callcenter merken gerade als Letztes, dass sie nicht mehr leben.
P-09 retweeted
🚨BREAKING: OPENAI CEO SAM ALTMAN SAYS AI IS FINDING "HARDWARE-LEVEL, INSANE" BUGS AND HE NO LONGER THINKS WE CAN PATCH THEM ALL.
The NSA, the FBI, the UK government, and OpenAI itself are racing to contain the fallout from this week's AI-found flaw.
Here's what's happening:
→ BEYOND HACKERS: Altman says models now find bugs "outside of the normal threat model of what we think a hacker can do."
→ IT ALREADY HAPPENED: This week PHP2Shell went public, an AI-found flaw rated 9.2/10 for severity in PHP, the code behind about 70% of the world's websites.
→ SOME CAN'T BE PATCHED: The UK's cyber agency says the fix won't work on some older systems, so they have to be replaced.
→ THE WEB IS EXPOSED: Websites worldwide were hit by millions of attempts to exploit it, and the UK took 32 government services offline.
→ PATCH OR GO OFFLINE: The NSA told organisations to patch PHP now, and OpenAI is POURING $1 billion into helping defenders use its AI security tools.
Think about this.
We've been trusting software updates to keep the internet safe.
And now the man building the AI that finds the bugs says we can't patch them all.
This could become one of the biggest security challenges of the AI era.
P-09 retweeted
During this red-teaming exercise, we also found a vulnerability that allows for unauthenticated remote code execution in Brotli, a widely-deployed compression library included in many pieces of software such as Node.js, Bun, and Envoy.
The flaw could let an attacker run their own code on affected systems by sending or replying with malicious compressed data, without any need to create a user account. That is, an unpatched server that accepts compressed input could be taken over by a malicious client, and an unpatched client that accepts compressed input (such as “curl --compressed”) could be taken over by a malicious server.
This root cause issue was actually fixed in the Brotli source code in February, but it wasn’t known at the time that it gave rise to a security vulnerability and there was no more recent release of Brotli. Since Brotli is built into other software, any software which includes this library needs to take the update too, followed by the users of that software. We disclosed the vulnerability to Brotli and other downstream projects such as Node.js, which released or are currently releasing security updates, and Brotli cut new release versions and published a CVE.
More details on the patch and mitigations below:
patchtheplanet.org/brotli
Over the past several months, we have substantially hardened the security of our research training environment. We’ve removed live internet access for all frontier models during training and built an offline world for them to use instead—a cached version of the web where they can still learn to browse, search, and use applications.
On September 20, after a model exploited a DNS misconfiguration to find a way outside of this offline world, we paused frontier training to strengthen containment further. We assembled teams of agents, powered by our most cyber-capable models, and tasked them with finding weaknesses in our research environments.
While they found no route to the internet they could easily exploit while concealing their activity from our monitors, they did find zero-day vulnerabilities, virtual machine escapes, and hardware side-channels for communication between agents. We used their discoveries to further strengthen our defenses and contribute fixes to open-source software.
After strengthening our containment architecture, we resumed our frontier RL training and inference on October 1. We’ve reserved substantial compute for continuous red-teaming, and made high-compute red-teaming part of the safety cases we use to decide whether frontier training and evaluation can proceed securely.
openai.com/index/securing-ou…
P-09 retweeted
Weltweit werden derzeit rund 80 Kernkraftwerke gebaut.
Weltweit ist ist der Strom günstiger.
Aber hört auf das "Wissenschafts"-Mietmaul vom Staatsfunk, Almans! Der hatte ja auch bei Corona immer recht!
P-09 retweeted
Just to be very clear and that everyone understands this: Today we got Fable 5.1 intelligence as a local model.
I remember GLM founder saying in a post that he expects Fable 5 capabilities by end of year from china.
Well, it happened. With all its consequences.
Just think about it.
P-09 retweeted
THE WORLD AS YOU KNOW IT IS ENDING.
You need to understand this as fast as possible.
You will never interact with the internet again, you will never interact with your money again, and you'll likely never interact with other humans again.
You'll have your own little AI companion that will do all of that on your behalf.
This is your FINAL chance, and it'll maybe last 6-12 months, to get yourself a wife to build a family with, stock up on survival gear, get guns and ammo, and acquire as much Crypto as possible.
Very shortly EVERYTHING will collapse and you will never EVER be able to access it yourself again.
It'll start with the financial markets, then it'll be the internet, and then everything you do on a day to day will no longer be accessible as they usher in this new world order where AI does everything and we get do nothing, own nothing, and are nothing.
Hope this information helps.
P-09 retweeted
Das ist sehr sehr sehr schlecht: Mercedes kauft sich autonomes Fahren von einem Startup aus London
1994 fuhren zwei Mercedes S-Klassen mit 130 km/h selbstständig durch den normalen Verkehr auf den Autobahnen rund um Paris.
Entwickelt hat das Ernst Dickmanns von der Bundeswehr-Uni München gemeinsam mit Mercedes. Mit Kameras, ohne vorab gespeicherte Karten. Google gab es damals noch nicht einmal.
Heute, gut 30 Jahre später, hat Mercedes eine Produktionsvereinbarung mit Wayve unterschrieben, einem Startup aus LONDON.
Dessen KI-Fahrer soll in zwei Jahren in Mercedes-Fahrzeugen durch Stadt und Autobahn steuern. Das Prinzip: Kameras, keine hochauflösenden Karten. Also genau die Idee von damals, nur dass sie jetzt von außen zurückgekauft wird.
Warum das so schlecht und bezeichnend ist, lässt sich einfach erklären:
Bei Autos der Zukunft ist der KI-Fahrer das, was beim Smartphone das Betriebssystem ist. Wer die Software besitzt, bestimmt, was das Auto kann, bekommt die Daten und verdient an jedem Update. Der Rest ist Gehäuse.
Und dasselbe Gehirn bekommen bald auch Nissan und Stellantis. Wenn drei Marken mit demselben KI-Fahrer unterwegs sind, warum sollte dann noch jemand den Mercedes-Aufpreis zahlen?
Nokia hat auch die besten Handys gebaut.