@JensenHuang@nvidia So, if hybrid models are the way forward, doesn't it follow that the infrastructure running them will become hybrid too? That's where i see huge potential for decentralized AI infrastructure, complementing centralized clouds rather than replacing them
Every decision looks obvious from your own point of view.
A company protects its platform, a government has to weigh privacy, freedom, security, crime and national interests at the same time. It won't always get the balance right but it can't ignore those responsibilities either.
@ionet Insane cost reduction and this perfectly shows the efficiency of DePIN, tapping into globally idle GPUs to let the market dictate the true cost of compute
@wojakcodes I think a major reason for this is a lack of legal awareness. If citizens don't know their legal rights then they can't protect themselves against abuse.
@devisopw@ionet@ionetindia@Gaurav_ionet well articulated bro, i agree & believe the hardware constraints are the new oil crisis of the tech world. Tapping into idle global compute via decentralized networks is one of the most practical ways moving forward.
AI is not becoming expensive because the models are getting smarter.
It’s becoming expensive because compute is not getting cheaper.
It’s one of the biggest reasons i am bullish on @ionet and the decentralized compute space.
From Uber exhausting its AI budget in just 4 months to the growing cost of running large scale AI…
Here’s why compute is becoming the biggest bottleneck for the AI industry and why decentralized compute could change that. 👇