Episode of cancel minutes day. I'm talking with lewis. Says he says who is a professor at the computer science and engineering school and u. dub as well as the founder and ceo of. Octo mel. how are you sir. Good good thank you think him. I'm looking forward to this executive in your show very kind so two full-time jobs. That's right. yeah you know. I it just takes so i promise. I'm not hyper distracted The work that leads to octa mouth started. The washington in the opus was the oprah's project called apache. Vm started there five years ago. When a lot of the folks that we've talked to the project to who found the company with me you know in Yeah it'd be spending most of my time to maladies days. So and this is the thing that's encouraged. This is actually your second spinoff from youtube Correct correct yet. We make mistakes. We learn to eat again. We do better. So that's that's encouraged when you're faculty ura you've got phd. You're doing postdoctoral work during research and you come up with an idea like it may be founded. Phd work may be funded by other companies. But then you can go in spin off a company i. How does that work do you keep track. He absolutely so yeah. It's it's encouraging you know. Companies around around research workers agreed tweets have impact know to you know when there's many ways have impacted research right so you write papers at the region following research sudas graduate and become professors into into the next thing. Or you know you went when the research is is a timely and you know has shorter debt has just prime for impacting the street today reform companies and luckily off. Timeout is extremely. Wow also very cool before. We get into october mel and apache. Tvm what they do. I wanna take a moment to back up about five years. Because i read a paper that she worked on With the computing community consortium that was called arch twenty thirty or architecture twenty thirty s a vision of computer architecture research over the next fifteen years and here. We are one third of the way into the predictions that you were making in this in this paper here which i think is the only twelve pages. It's worth reading. I think it'd be useful to talk about those things and see how close you are because Machine learning was a big part of that. Yeah absolutely so. Yeah that that. That was a really really interesting fun to write. Was the result of workshop that i- koran when my colleague tom winners and mark hill And you know the way i see compare conductors. That's my home community in computer in in computer science computer sectors really about how to build systems. How'd you put harder and softer together to enable.