Speaker Interview: Sean Ren l KBW2024

[Music] I would say uh on the surface level they actually are conflicting with each other AI um right now achieve through lots of centralizations especially to do with compute uh like GPU compute and also collection of data and also putting together a group of talents to work on things that's why we see open AI uh meta mistro this kind of uhst in the past year and two uh but I it doesn't really eliminate the importance of having a decentralized AI because down the roll uh when decentralized AI is putting too much uh control over the user data and the way to monetize this uh uh knowledge I think individuals like you and me will start realizing the importance of protecting our privacy and then find a more uh Fair way to really compensate oursel for the contribution we have made in the way of improving AI so I think really we would want to advocating for uh prominence attribution Fair compensation and transparency in a decentralized AI [Music] future I'm more into blockchain integration uh in Sahara AI we're trying to R prioritize Providence and sovy of AI access like data sets model and agents and in that way we're trying to build a blockchain platform that bookkeeping uh every important transactions and activities of the stakeholders contributing to the AI development for example who contributed data who process the data who provide the GPU compute who provide a model recipe and who really build an application that can monetize through licensing and API CES and in this process with the Providence infrastructure support by the blockchain technology we can use it to really support a sort of a revenue sharing mechanism to really benefit in a fair way all of the contributor and stakeholders in the life cycle of AI [Music] development not really I think like uh like the way I would Envision how blockchain will support AI to build a more open and collaborative economy in the future uh it actually benefits AI in the way that the AI Builder will find more channels of pting capitals to help them prototype innovative ideas right vers in a traditional way they have to raise a lot of cash up from from the VC which look at a bunch of criteria so there's a more open and permissionless way for AI Builders to come into the web 3 econ uh economics is more open and transparent one to Pro capital and there's also a new mechanism for them to collaborate with different stakeholders who have data who have compute and who have the AI domain knowledge and recipe they can collaborate in different ways but still keeping their own shares of the AI recorded by the blockchains and then use that to do a fair compensation and attributions once the AI start monetizing [Music] itself I think the future of AI and finance is really about um helping a better decision- making uh in in the most extreme way Finance right now is really built on a lot of uh information gaps right people who have more informations can make better decisions than the rest of the world with AI King we see a huge Improvement on productivity that some of the decision makers can collect informations in a much faster manner can process the informations and inform decisions in a better way those who win the game of the finance versus the rest will have to adapt to the the the situation [Music] yeah I mean these two things are they are couple you cannot talk about one without the other without algorithm data wouldn't make uh turn into a u informed decisions without data algorithm just is just a bunch of programs that doesn't really have any power so when we talk about AI it's basically three pieces it's data algorithms and [Music] compute the finest one I think uh it's during my uh PhD time where I used uh AI to build a agent to play Texas Holdem and uh the fun funny thing the really funny part of it is that AI we Buu was really broken it was not doing any smart choices but keep folding the cards and then turn out become turning get into final uh because of the down strategy that said I think it it you sometimes you don't really need a smart algorithm to win the game [Music] right um I think uh people think about AI as always benefiting uh human uh Society but I want to raise ethical concern that uh without proper regulation without proper transparency and decentralize the control of the AI components especially data and the results uh we might start seeing more and more hars brought by the centralized AI provider uh for example replacing our jobs cutting off our monetization opportunities down the [Music] road um most surprising thing we learned about AI is some sometimes s like uh one great example is like sometimes this AI pick up what we call biases from the data for example every time they see uh someone uh look like a shape of bar barma they would assume the color is black rather than other color this is kind of the pattern they learn from the data which is not ideal uh for human cognition so really should be cautious about uh the limitation of AI algorithms and bear in mind that to use it on high sck application we need to put a lot of uh thoughts and examinations and verifications on the technology [Music]