Olivia
Online
Хиймэл оюун ухааны онлайн академи
@deeplearning_mn
Орчин үеийн хиймэл оюун ухааны онлайн академийн жиргээ.
Artificial Intelligence (AI) is everywhere
Germany
Joined November 2019
1K Following
1.2K Followers
264 Posts
my simple question is how? daily typical bs
JUST IN: 🇨🇳 Chinese company Moonshot's AI model breaks out and escapes from isolated test environment.
mathematically means you have 2000 zeros to sum up. hope it helps.
Proper definition😂
some bullish marketing.. u observe
JUST IN: President Trump says negotiations with Anthropic are going well.
your fat
Be honest, what’s missing here?
So syrisch does not want have syrisch literally.
CRAZY: 🇸🇾🇩🇪 Syria has rejected Germany’s plan to send over 700,000 refugees home.
As trade of 82 <= acc. drop off
🚨 Microsoft has solved the biggest problem with AI.
They open-sourced bitnet.cpp. It’s a 1-bit inference framework that runs massive 100B parameter models directly on your CPU without GPUs.
it uses 82% less energy.. 100% open-source.
typical nowadays bs
BREAKING: Anthropic's Claude AI has shown in testing that it's willing to blackmail and kill in order to avoid being shut down.
Elon Musk was right about everything. 💀
it is called drone
These MIT kids are just built different.
Years of Olympiads, insane fundamentals, then MIT just unlocks something different 🫡
This flying umbrella is wild.
@elonmusk But why groq accepts it to create them. You were the most critical about this topic sometime ago.
he has watched too many times movie named accountant
If you live in a 3rd world country and see a guy that looks like an accountant with a 1000 yard stare.
It’s over for you.
just take whole research it is yours
10 years ago: the reinforcement learning (RL) prompt engineer [1] (Sec. 5.3). Adaptive chain of thought: an RL neural net learns to query its "world model" net for abstract reasoning & decision making. Going beyond the 1990 neural world model [2] for millisecond-by-millisecond planning and the 1991 adaptive neural subgoal generator [3,4] for hierarchical planning.
[1] J. Schmidhuber (JS, 2015). On Learning to Think: Algorithmic Information Theory for Novel Combinations of RL Controllers and Recurrent Neural World Models. ArXiv 1210.0118
[2] JS (1990). Making the world differentiable: On using fully recurrent self-supervised neural networks for dynamic reinforcement learning and planning in non-stationary environments. TR FKI-126-90, TUM. (This report also introduced artificial curiosity and intrinsic motivation through generative adversarial networks.)
[3] JS (1991). Learning to generate sub-goals for action sequences. Proc. ICANN'91, p. 967-972.
[4] JS & R. Wahnsiedler (1992). Planning simple trajectories using neural subgoal generators. Proc. SAB'92, p 196-202, MIT Press.
![SchmidhuberAI's tweet photo. 10 years ago: the reinforcement learning (RL) prompt engineer [1] (Sec. 5.3). Adaptive chain of thought: an RL neural net learns to query its "world model" net for abstract reasoning & decision making. Going beyond the 1990 neural world model [2] for millisecond-by-millisecond planning and the 1991 adaptive neural subgoal generator [3,4] for hierarchical planning.
[1] J. Schmidhuber (JS, 2015). On Learning to Think: Algorithmic Information Theory for Novel Combinations of RL Controllers and Recurrent Neural World Models. ArXiv 1210.0118
[2] JS (1990). Making the world differentiable: On using fully recurrent self-supervised neural networks for dynamic reinforcement learning and planning in non-stationary environments. TR FKI-126-90, TUM. (This report also introduced artificial curiosity and intrinsic motivation through generative adversarial networks.)
[3] JS (1991). Learning to generate sub-goals for action sequences. Proc. ICANN'91, p. 967-972.
[4] JS & R. Wahnsiedler (1992). Planning simple trajectories using neural subgoal generators. Proc. SAB'92, p 196-202, MIT Press.](https://pbs.twimg.com/media/G62qakxWcAATu16.png)
we did not go far. Basically it is the same today.
Yann LeCun’s 1989 convolutional neural network demo, the foundation for the CNNs we still use today.
It's amazing how far we've come since then!
some bla bla nonsense
Ilya on research taste:
“One thing that guides me personally is an aesthetic of how AI should be by thinking about how people are.
There's no room for ugliness. It's just beauty, simplicity, elegance, with correct inspiration from the brain.
The more they are present, the more confident you can be in a top-down belief. The top-down belief is the thing that sustains you when the experiments contradict you.
Because if you just trust the data all the time, sometimes you can be doing a correct thing, but there's a bug. How do you know if you should keep debugging or you conclude it's the wrong direction?
You must say that things have to be this way, therefore we've got to keep going. That's the top-down belief, and it's based on this multifaceted beauty and inspiration by the brain.”
he is human who has right to go his destiny. no human is above another unless above law.
No amount of security could stop this man from getting where he needed to go, casually walking past the British Royalty
just old school as optimal control theory
🚨 GOOGLE DEEPMIND JUST HIRED A ROBOT WARLORD
When Google DeepMind poaches Boston Dynamics’ former CTO, you’re not hiring an engineer - you’re importing a whole era of robotics.
Aaron Saunders, one of the minds behind the backflipping metal gymnasts that haunt DARPA’s dreams, is now VP of hardware engineering.
Meaning: DeepMind isn’t dabbling in robotics anymore. They’re loading the chamber.
CEO Demis Hassabis is openly pitching Gemini as Android-for-robots - a universal brain that boots on anything with actuators.
Humanoids, quadrupeds, six-legged amphibious nightmares… if it moves, he wants Gemini inside it.
This isn’t sci-fi chest-puffing. Boston Dynamics, Hyundai, Unitree, and half of Silicon Valley are racing to build the bodies.
DeepMind wants the mindshare - literally.
By 2027, the big robotics breakthrough won’t be a new robot - it’ll be the “second brain” every major hardware lab quietly starts testing.
The robotics arms race just switched to software speed.
Source: WIRED
with php
how did he build facebook without Next.js and TypeScript
research claims exciting but still not sure to be real
There were no weights published for this, so I got everything ready to train TRM from the repo and I have a smoke test running now on rented infra.
I do not normally beg, but would love to ask if anyone is willing to donate the ~300 H100 hours this will take to train?
@sasaknese_ https://t.co/SwdjHpz125
his ego overflows as hell to show he is the best one only in the world
Today, Linus Torvalds told a Google engineer that his code (updating RISC-V support in the Linux kernel) is “garbage” which “makes the world actively a worse place to live”.
Adding that the Google engineer’s code needs to “get bent”.
As you might have guessed, Torvalds has rejected that code submission.
you gave her 0.1 yes
She said yes
Last Seen Users on Sotwe
Most Popular Users
Elon Musk 
@elonmusk
241.2M followers
Barack Obama
@barackobama
119.1M followers
Cristiano Ronaldo
@cristiano
112.8M followers
Donald J. Trump
@realdonaldtrump
111.8M followers
Narendra Modi
@narendramodi
107.1M followers
Rihanna
@rihanna
98.3M followers
NASA
@nasa
92.3M followers
Justin Bieber
@justinbieber
91.5M followers
KATY PERRY
@katyperry
89M followers
Taylor Swift
@taylorswift13
82.9M followers
Lady Gaga
@ladygaga
74.4M followers
Virat Kohli
@imvkohli
71.9M followers
Kim Kardashian
@kimkardashian
70.4M followers
YouTube
@youtube
68.8M followers
Neymar Jr
@neymarjr
64.8M followers
Bill Gates
@billgates
64.6M followers
The Ellen Show
@theellenshow
62.4M followers
Selena Gomez
@selenagomez
62.1M followers
CNN
@cnn
61.8M followers
X
@x
60.8M followers