@balajis
Decentralization is the way. This is why we
@opentensor
have been working on the Bittensor protocol: a decentralized network of AI models, all competing with each other to provide the best possible output, enabling AI training and inference that is continuously improving.
By way of an update, we have contained the attack and put the chain into safe mode (blocks producing but no transactions are permitted).
We’re still mid investigation and are considering all possibilities. Stay tuned.
New Podcast 🎧
Ala Shaabana: Beyond the Silicon Valley Echo Chamber: Decentralizing AI with
@opentensor
Open vs Closed AI systems will be one of the most important convos in our lifetimes
Apple:
Spotify:
Takes by
@shibshib89
👇🏻
This is exactly why AI can't be controlled by a select few people. Corruption, squabbles, pursuit of profit despite promising "open ai", all of this in one self-appointed group. Scary thought.
@neural_internet
@bittensor_
Well said! Another magic of Bittensor is it's ability to swallow *any* technology. This allows anyone with a model to join the network and work with the API.
@coverageproved
@opentensor
@bittensor_
@Old_Samster
Open source AI will always be more horizontally scalable due to the capabilities of decentralized AI projects like
@bittensor_
, whereas closed source will always require vertical scaling, an inefficient process that is slower and much more expensive than horizontal scaling.
@JosephJacks_
@KeithSingery
Subnets should be treated like startups, start off by constraining the scope to a single problem. (Say, for sport betting, on a single sport/given event). Solve that problem. Expand the scope afterwards. Build like a startup, and you will succeed.
We have Ala Shaabana on the Infinite ML pod today
He's the cofounder of Bittensor, an open-source protocol that powers a decentralized machine learning network. He has a PhD in computer science and was previously at Instacart and VMware.
In this episode, we talk about:
@FalaahArifKhan
I remember a while ago some folks wanted to prove these journals are bogus so they wrote a text generator and just "generated" a paper (this was in mid 2000s, before the rise of deep learning).
Fake paper was published of course.
@XRPeeved
@unconst1
You're right! People have said this so many times over the years.. But I think what Bostrom was trying to say is that once we achieve artificial super intelligence it'll be able to do everything we do and more, such that it’ll be able to create everything else there is to create
@ModernEremite
Hey
@ModernEremite
! I had meant that we are reaching GPT-2 in collective size, in fact I somewhat misspoke as we have surpassed its size at this point. We are now working on validation and benchmarking to find out just how close we are to state of the art (SOTA) performance!
@zachklipp
Mutations are bound to happen. Vaccines account for them and tweaking vaccines to radical mutations isn't as time consuming as making a new vaccine.. All these mutation reports are sensationalism, nothing more
@ModernEremite
Knock on my twitter anytime friend! You can also join our discord where you’ll likely get a faster response since all the devs are there.