ai.bythebay.io: Adam Gibson Interview

[Music] I just find AI is actually way too broad I just find AI is actually way too broad and it's actually it's actually several and it's actually it's actually several subfields so when people think about AI subfields so when people think about AI they should think about the first thing they should think about the first thing should ask is what field and by default should ask is what field and by default you should probably think machine you should probably think machine learning and then in the context of my learning and then in the context of my work we're departing company so work we're departing company so everything everything we do in AI is everything everything we do in AI is machine intelligence related deep machine intelligence related deep learning related right so a lot well so learning related right so a lot well so a lot of honestly a lot of what works a lot of honestly a lot of what works right now is we can count things and we right now is we can count things and we can we can point at things so we can can we can point at things so we can count and we can label that's what works count and we can label that's what works really well right now and that's what really well right now and that's what that's what most people are actually that's what most people are actually deploying you might and then you may be deploying you might and then you may be you might add forecasting in there you might add forecasting in there everything else doesn't so if you hear everything else doesn't so if you hear about some new paper from Google it's about some new paper from Google it's probably not gonna work for you because probably not gonna work for you because it's actually just it's a bleeding edge it's actually just it's a bleeding edge paper meant to demonstrate a concept paper meant to demonstrate a concept like most AI out there that you see is like most AI out there that you see is that's really cool that's really cool is actually just a proof of concept for is actually just a proof of concept for something that's gonna be built on over something that's gonna be built on over time and then within with it so within time and then within with it so within my subfield it turns out that's how we my subfield it turns out that's how we make money we label and count things and make money we label and count things and we do forecasting and then from there it we do forecasting and then from there it just depends on the vertical that you're just depends on the vertical that you're dealing with um honestly like actually the hype the um honestly like actually the hype the hut you know cuz because you know hut you know cuz because you know because we build deep learning for the because we build deep learning for the fortune 2000 and a lot of those guys fortune 2000 and a lot of those guys don't care what the algorithm is they don't care what the algorithm is they don't care what the math is they just don't care what the math is they just they see the hype and they're just like they see the hype and they're just like it's magic right you can do anything I'm it's magic right you can do anything I'm like no no no so there's a lot of like no no no so there's a lot of problems here and then half of what I do problems here and then half of what I do is market education so if I could you is market education so if I could you know the other thing is you know data know the other thing is you know data scientists they you know they think scientists they you know they think because they can spin up AWS and they because they can spin up AWS and they can run a tutorial that they're deep can run a tutorial that they're deep learning expert no you're not you have a learning expert no you're not you have a lot of work to do like neural nets lot of work to do like neural nets themselves are empirical you need to themselves are empirical you need to tune them you need to spend time tune them you need to spend time actually trying to understand what does actually trying to understand what does this graph tell me you get you need to this graph tell me you get you need to do that if you're gonna do like if do that if you're gonna do like if you're gonna succeed in this field it's you're gonna succeed in this field it's not magic and you shouldn't treat it as not magic and you shouldn't treat it as such so I so I mean I've been a speaker such so I so I mean I've been a speaker to Lexi's conferences for last couple to Lexi's conferences for last couple years now and I always love how years now and I always love how down-to-earth and practical the down-to-earth and practical the conference's are you know it's great conference's are you know it's great networking he gets great speakers in the networking he gets great speakers in the room he always picks amazing venues and room he always picks amazing venues and then oh you know so the networking is then oh you know so the networking is quality you know the you know it this quality you know the you know it this the conference is small which is amazing the conference is small which is amazing like it's it's just the right size for like it's it's just the right size for you to get networking in and and all you to get networking in and and all that other stuff so it's it's just a sum that other stuff so it's it's just a sum of a lot of small things of a lot of small things you