Subscribed like a few min in... You nailed it on the head and respect for the podcast too. I'm sick of these AI videos talking in monotone about basic level content for nearly every topic. AI in healthcare is interesting - I work in mental health and my company works at a national level for suicide prevention and we are rolling out in a couple ways; including for quality assurance of recorded conversations to identify statements that match identifiers (Lyssn). There may be other ways leadership is looking to integrate AI. I just started at a local Intel/Dell AI Workforce program and hope to branch into AI. I live fairly close to Oregon State University and their AI program has a massive investment from NVIDIA founders for a dedicated research building. Really excited to see where this field takes me. Thanks for the inspiring work here!
Awesome video! Totally agree. I think one thing you didn't cover but will likely change is movies and TV shows. I think there will be several developments. I think, like you mentioned, barriers to entry will drop so low budget movies will be able to look amazing. This will lead to lots of niche content with strong followings. I also think that big budget movies will use AI to let people change movies to suit audiences better. This will start with something basic, like using AI to create a child-friendly version of a movie where nudity, violence and profanity is taken out. It'll also be used for dubbing. But it will go crazy when people start watching movies where they deep fake their faces onto it so they can watch the Avengers but it's you, your mates and your families in the movie and the characters actually look, sound and maybe even act like you. I also think that content creators will start building out the extended universe of TV shows like Star Wars. These won't be canon but it makes sense to let people take the IP and make their own fan pics and then pay a cut to the corporations that own it. There will be restrictions about how they can be used (i.e. no Star Wars porn parodies using AI to use the original cast) but it seems inevitable and makes business sense as the best stuff will make them money and everything else costs nothing. They can probably remake the best content and not even pay the original creators if they find something is working. Basically crowdsourcing blockbusters (I don't think that would be fair but that is very separate from whether I think it'll happen).
although i'm currently a cybersecurity student, i've noticed that field already utilizes AI for detecting and mitigating threats as two examples. i do wonder how much more things will augment as time goes on. c: this is probably an unpopular opinion; however, i find it rather oddly fascinating to live during this time, especially as a sci-fi lover. i compare the feeling to that of: "pinch me, am i dreaming?"
I think disruption implies that it will change how they fundamentally do business. I think this isn't bad for the consumer at all, but it probably means that a lot of companies operating in this space will have to adapt or close down!
Does the idea of training an AI using AI-generated data seem super sus to anyone else in that we're potentially building a garbage-in-garbage-out feedback loop?
ChatGPT was trained on the general internet, inter alia. Do you know how much garbage must of gone in? They do human feedback training to make it prefer high quality solutions so in a way, GIGO no longer applies to modern AI.
@@michaelnurse9089 You're talking about RLHF. I'm talking about training that takes humans out of the equation almost entirely -- like how the people hired to generate training data are using AI to do it. The general Internet data used to train ChatGPT was still of human origin.
Realistically, I think it will decrease these jobs in the long run (5+ years). I think you can future proof by learning how to use the AI tools and build products. I see most ds roles moving more towards product ownership. I think this skillset will be increasingly useful in the long run
6:43 Crazy how a press release surges intrest in a tool. Especially in this Copilot case. It's such an undervalued tool right now and X is just going to take it to another level
What your undergrad majors in? I am second year CS and stats student thinking of doing master in applied statistics (granted my grades good enough) to become statistician or biostatistician. I would say it doesn't matter too much between applied stats and biostatistics, however applied statistics is more versatile whereas biostatistics is more specific (meaning you would be better candidate for biostatistician type of roles). If you want to be a biostatistician I would say choose biostatistics but if you are unsure like me choose something more generic. But of course I am a random stranger so choose what you want after research, don't do something just because one person said to do it, whatever you choose I'm sure in the long run it would aid your career.
I think it has similar challenges to medicine where there are a lot of regulations. Still, some fintech startups will also probably disrupt that industry!
It is not AI that will cause the changes. It is the way people will use it. Human to human contact is one of the great drivers of constructive activity, when this decreases humanity will increasingly become separate from each other. I'm not sure of the consequences but I am sure it will not be be helpful.