Episode Details
Title | The No BS Guide to Crypto AI with Prodia - Shawn Wilkinson, Prodia, Ep. 260 |
---|---|
Date | 8/14/2024 |
Podcast | The Blockcrunch Podcast β |
Key Insights
Bullish:
- π’ $Protea - Discussing the potential of Protea's future developments, cost efficiency, scalability, and expertise in distributed systems.
- βWe can just distribute that to our network. Oh, the cost? No problem. We'll cut that in half. And, oh, by the way, it's like, you know, 2 to 4 times faster.β - Shawn (0:00)
- βWe were able to scale that product out, and then we kind of realized, oh, wait. This is 50 to 90% cheaper. It's 2 to 4 times faster, and it's way more scalable.β - Shawn (0:00)
- βWe have all the experience in terms of building out distributed systems. You spend 10 plus years, myself, building out distributed storage. We have all the expertise to build out distributed compute.β - Shawn (0:00)
- βWe literally went from, you know, a cold outreach to, you know, a done deal. Very, very small. Yep. Cold outreach, small amount of time, and just because the pain point was was just, like, so there.β - Shawn (0:00)
- βWe have a highly simplified cost structure. Basically, if you look on a proteo.com/network, you see our current pricing for different GPUs. We just pay a fixed amount per GPU per month.β - Shawn (0:00)
- βWe start from the other. So we don't focus on the supply side. We focus on demand side. So our utilization rate is 100%. We use 100% of the GPUs, that are available in our network.β - Shawn (0:00)
- βWe are able to solve their problems, but, yeah, where our goal is to to help, many, many, many people as they try to build things in AI.β - Shawn (0:00)π
- βWe're probably gonna be, you know, as we grow, probably the largest user of all these decentralized networks.β - Shawn (0:00)π
- βWe have more usage than any other individual network, and probably in the next, you know, 2 months or so will be larger than all of them combined in terms of usage.β - Shawn (0:00)π
- βWe just wanna make it simple. You know, just like a product, we want to, you know, integrate in 5 minutes. We want to make the supply side initially as simple as possible.β - Shawn (0:00)π
- βWe think AI is gonna be in everything if it isn't already, in terms in the title, but we think it's it's gonna power, you know, a lot of things.β - Shawn (0:00)π
- βWe have partners that we can help cut connect cost customers with that can, you know, get the back, you know, backroom deal for for the high end GPUs.β - Shawn (0:00)π
- βWe are able to sit down and write, you know, different acceleration libraries and things that like an individual, you know, company or team, it wouldn't make sense for them to do themselves.β - Shawn (0:00)π
- βWe have access to all these additional models and features that you don't have, right?β - Shawn (0:00)π
- βWe have the right hardware in the right locations, right latency, right connections, and in all, it kinda sums up to to those numbers that I talk about.β - Shawn (0:00)π
- βWe have a version 2 of our API, coming out for Stable Diffusion Excel workloads. And it's pretty much the fastest in the business at about 2 seconds.β - Shawn (0:00)π
- βWe power some of the largest, essentially AI, image editors and apps out there. Literally, if you Google, you know, AI image editor, you know, the number one result there, 70,000,000 users uses us behind the scenes.β - Shawn (0:00)π
- βWe have all the expertise to build out distributed compute. Let's do that.β - Shawn (0:00)π
π More quotes available for premium users