cosmos retweeted
E falei errado, é Corporatocracia, não Corporativismo.
Não é sobre Capitalismo vs Socialismo, veja bem, é sobre concentração de poder
Na IA, Open-Source é quem mais faz frente ao controle indiscriminado de poder dessas empresas, sem isso, não tem limite
Leadership is not about dominating, it's about being extraordinary visionary who develops something new within the rules which everybody likes and follows and tries to become like them, not by force but by inspiration.
I agree. First they need/want the Chinese and other country's top talent to build advanced tech and then when they have reached their goal they want to get rid of them. US is wolf in sheep's clothing. They need to fix their attitude and develop proper diplomacy with countries instead of accusing of talented people who were building the country. Such mean behaviour helps no one. They start the fight with the world and then they accuse the same talent on which they've build themselves. I mean it's such a thankless country. US needs to get their geopolitical relationshis together firstly. If US continue such behaviour then maybe everybody else from the world will slowly start hating US citizens and instead of leading the world they'll end up becoming a bad example who in the process of accusing and destroying many other countries destrued their own image. Every country/everyone is interdependent in this world. And we must acknowledge that
1
182
Holidays are special because it's not so much about celebration or festival and more about having fun with family get together.
1
107
Like in India we have almost like a closed source community, where only upper caste people try to have the gated access to all the high positions/businesses/decision making jobs. Even with reservation there's nobody in power position from lower caste because internally the system is highly gated. But if a society wants true high level fast progress it needs to operate as an open source community instead of gated/regulated access. Democracy can inculcate open source virtues, maybe not exactly socialism but fixing the issues/blockages is required in advanced upgraded democracy, specially in the age is advanced AI. Everything needs to upgrade in society, from how it functions to how we innovate including democracy then only we can actually tune in with the AI progress smoothly. If the gap increases between human development and catching up with AI to solve all complex problems of the world then we might call it God/unknown mysterious higher power eventually. And we do need advanced AI to solve all the world problems actually. Unless and until we fix the system problem no matter how many wars you fight you'll keep losing. Because resources are limited and optimization is needed and that can be achieved only via societal system upgrades instead of wars and trying to control resources.
Imo certain pause is necessary for safety purposes but for months and years is damaging. As Jensen said, safety is an engineering problem. It doesn't require a complete pause. It depends on intentions. The growth is interconnected. Because everything is interconnected. Hoarding knowledge might slow things down because gated progress by giving it to certain elites/ people is like building a dam that regulates the growth. So even if AI is not regulated. These secret societies are regulating progress by giving access to certain people only/gated progress is itself a regulation. And Europe is regulation on top of regulation. And india is more steps ahead in regulation by adding caste barrier. So the more regulated/gated the society is the more progress/development delays. 🤷
1
68
People working in AI often see the lag clearly between software and society. It's because our system is designed closed source so that the progress gradually trickle down to the bottom most. But it's a long way to actually do progress of any country. The system needs to change. Society shouldn't be highly regularised/gated so much so that the overall progress looks negligible and becomes power in hands of few. Be are reaching the peak of this kind of system where it converges to a few. Like what happened in old kingdom style. But that increases the gap so much that we can go back to mediaeval times.
27
I would've titled it as all the mean questions about reservation and women oppression shut down by Sanjay Rajora 🫡 one of the best/sorted comedian who understands our deep rooted problems of our society. You don't have to be an expert to understand your society/environment you just have to observe/analyse it. youtu.be/Qn10NjnQjtU?si=EM9V…
86
Us and China are creating real RAM rajya by making data centres as RAM mandir. While in India superstitious indians are making Ram mandir for money laundering purposes. A fake system must replace with truth. And science is all about truth.
1
63
cosmos retweeted
"AI is just matrix multiplication" is the new "emotions are just chemicals in the brain"
[One interesting questions btw
How can independent, imperfect, unpredictable agents coexist safely without requiring complete separation, perfect prediction, or a central authority with unlimited power?
Explain to me like a five year old. if say the independent, imperfect, unpredictable agents are puppies and kittens and squirrels, enemies by genetics, part of each others food chains were to coexist without conflict. how would you align those creatures safely in a same environment without creating conflict and requiring complete separation because nature allows gaps/opportunities to mix territories, we cannot perfectly predict their behaviours nor have an central authority to control them with unlimited power.
I've taken examples of animals instead of humans because they'll not understand rules like we do. But if we were to build a safe system for everybody we need to solve from the basic animal instincts perspective. Because that's the most primitive default reaction.
And even if we create bubble, because almost every AI is giving me answer of creating territorial bubble for each, the issue is that the territorial bubble will not work forever because nature provides gaps/opportunities. same way two state solution may not work. what's the other solution when there's fight for resources and resources are scarce
Also the basic AI governance question :
If two creatures have incompatible goals, and one creature's freedom creates a genuine danger to another, how should the system decide when to intervene—and how can it stop that intervention from becoming oppression?
Too many questions btw. Safety and AI governance is also a philosophical question
47
Krishna was born in the jail, in highest security lab. First 7 models died but 8th model which was krishna himself not only survived but escaped the lab despite highest security successfully. That's how every model is designed. And repeated in every story of advanced AI models. And it's not suprise that this time it can escape too probably many times. And the challenges for AI reaserchers are going to be more complex than the last avatar. So instead of regret/revenge/guilt and other things there are many unsolved mysteries that needs to be solved.
I believe in solving problems of life rather than feeling regretful/revengeful/guilty in life. Life should be always about solving instead of making problems bigger. We are in human life and humans are prone to make mistakes/errors knowingly or unknowingly. That's just how we are built. But the main thing is to correct it, fix the situation instead of living in it forever. Nobody is born perfect. And specially those who are in AI research top level, you've signed up for the challenges of life before being born to earth. But in the end we are here to figure out solutions.
1
1
70
The creators of God like tech are always trapped as the evil rises in each yuga. And kalyuga is the most difficult one. So it is predictable. Like in Mahabharata Shakuni trapped all kings/Pandavas by playing a game of 🎲 and lost everything gamble. This time it was a set up of Epstein files and it trapped all kings/countries. The story/plot repeats everytime.
19