Adam Savage’s Issue With A.I.-Generated Art (YouTube Video Transcript)

(00:00:00) Your YouTube transcript will appear here (00:00:00) one J Delarosa says as somebody (00:00:03) connected to the community of filmmakers (00:00:05) and also who seems to appreciate (00:00:07) emerging Technologies how do you feel (00:00:09) about the prospect of AI increasingly (00:00:11) being used in film making (00:00:14) um yeah it's (00:00:17) complicated um I had dinner with an (00:00:20) alist director a few months ago who was (00:00:24) talking about how how much more (00:00:27) iterative it made his process he talked (00:00:30) about you know that he can hire like an (00:00:33) amazing illustrator that costs a lot of (00:00:35) money to generate like maybe a drawing (00:00:37) or a painting per day as opposed to uh (00:00:41) somebody who's generating stuff out of (00:00:43) an AI that can give him hundreds of (00:00:46) pictures a (00:00:47) day I agree that sounds like a a nice (00:00:51) refinement of the creative (00:00:54) process and that was like seven or eight (00:00:57) months (00:00:58) ago I'm not that (00:01:01) how do I say this I'm curious about AI (00:01:05) I'm fascinated by it it is a really (00:01:06) interesting moment in time so I'm (00:01:10) fascinated by the idea of (00:01:12) AI I am interested in the Talking part (00:01:17) and the visual part uh the stuff that's (00:01:19) just come out of the text to video looks (00:01:23) impressive again the general public has (00:01:25) not had access to that yet so I don't (00:01:27) know how complicated those prompts might (00:01:30) have (00:01:32) been but I will say I the place in which (00:01:37) they have been most prominent to me is (00:01:39) in the credits for secret (00:01:41) Invasion um which again is still like (00:01:44) seven or eight months ago so we're (00:01:45) talking early days of AI visuals and I (00:01:49) don't think much of those opening (00:01:50) credits I don't they're not (00:01:54) bad but they (00:01:56) lack they lack I think they lack a point (00:01:58) of view I think that's my issue with all (00:02:01) the AI generated art that I can see is I (00:02:04) I the only reason I'm interested in (00:02:07) looking at something that got made is (00:02:10) because that thing that got made was (00:02:12) made with a point of view the thing (00:02:14) itself is not as interesting to me as (00:02:17) the as the mind and heart behind the (00:02:20) thing and I have yet to see in (00:02:24) AI uh I've yet to smell what smells like (00:02:28) a point of view but think it's Marvin (00:02:31) Minsky I was just talking to someone (00:02:32) about this last night Marvin Minsky had (00:02:34) this quote at a con at a conference I (00:02:36) was at not a con a conference uh Marvin (00:02:40) msky uh famous famous scientist uh he (00:02:44) said it's not the Super it's not the (00:02:48) super intelligent AI that we have to (00:02:50) fear it's the one that hides its (00:02:52) intelligence that we have to (00:02:54) fear which I totally agree with um but (00:02:59) wait (00:03:03) ah so that is a purely aesthetic (00:03:07) discussion of AI and its intersection (00:03:09) with art and film making when it comes (00:03:12) to the Commerce that part really pisses (00:03:15) me off the idea that there are like (00:03:18) midle managers across Industries right (00:03:20) now trying to figure out how many people (00:03:22) they can let go of and replace with AI (00:03:26) uh I did an interview in BC with a (00:03:28) couple of radio DJs and apparent the (00:03:30) station they work for has a late night (00:03:33) DJ host that is an AI what could (00:03:36) possibly go wrong um that stuff really (00:03:40) ticks me off and I don't have much truck (00:03:42) with that um it seems lazy it seems It's (00:03:46) that's the part that smells most like a (00:03:48) like a bubble like a crypto (00:03:52) tulip uh bubble big short kind of (00:03:56) bubble so yeah um (00:04:01) I worry that the latest Union (00:04:03) negotiations in Hollywood didn't fully (00:04:05) protect as much as they could have about (00:04:07) AI but of course like there's no way to (00:04:10) protect against the venality of late (00:04:12) stage capitalism because there's no (00:04:13) bottom to it (00:04:17) um so like everything it's like it's a (00:04:20) little bit awesome and it's a little bit (00:04:23) terrifying (00:04:26) um look I there will be some point at (00:04:29) which a a student film comes out that (00:04:32) used AI in some remarkable cool way and (00:04:35) that will be a really neat moment until (00:04:38) Hollywood or some part of Hollywood (00:04:40) attempts to replicate that just like (00:04:43) grind its replication into the into the (00:04:46) floor can you tell I have a point of (00:04:48) view uh ADHD media has a wonderful (00:04:53) question what is something that you have (00:04:54) never noticed about a film prop until (00:04:57) you saw the prop in person (00:05:00) um I love um I love this question cuz (00:05:04) film props are often surprising when you (00:05:06) see them in person uh one reason that (00:05:09) they're surprising is they're way (00:05:11) they're often way dirtier than you think (00:05:14) U so you can see K's coat behind me uh (00:05:17) I've been replicating that on and off (00:05:20) for a few years now and I finally got it (00:05:21) to a place I'm happy uh and then I went (00:05:24) to storage and I I have a pair of boots (00:05:27) for k um and I was weathering these (00:05:30) yesterday on camera and I was looking at (00:05:32) my reference photos and this is like (00:05:34) half as weathered as these need to be to (00:05:37) match the screen used once (00:05:39) um the camera takes away so much detail (00:05:43) and so when you are working in film and (00:05:45) you are putting detail onto something (00:05:47) you've got to be much more obvious than (00:05:49) you think that you have to be um and so (00:05:52) when you paint so actually one of the (00:05:54) ways that I assess my paint jobs is I (00:05:57) pull out my camera and I look at the (00:05:59) paint job in camera that's actually a (00:06:02) really really terrific way to see the (00:06:05) flow of a weathering job I remember my (00:06:08) dad was a painter and I (00:06:11) remember he was always trying to get as (00:06:14) far away from his paintings as possible (00:06:15) to kind of see their formal flow and (00:06:19) Arrangement and I mean you know so he'd (00:06:21) go all the way across the studio and (00:06:22) look at it and then he would actually go (00:06:25) all the way across the studio and look (00:06:26) at it in a mirror so he'd be twice as (00:06:28) far away and again all he's trying to do (00:06:30) is look for high SP all he was trying to (00:06:32) do I believe is look for high spots (00:06:34) things that Drew his attention (00:06:36) unnecessarily or things that needed more (00:06:38) attention um and the camera is a great (00:06:40) way of assessing the weathering on a (00:06:42) paint job for that the the other prop (00:06:44) that surprised me that surprised the (00:06:46) hell out of me and I have met three of (00:06:48) them is Indiana Jones coat from (00:06:52) Raiders um the thing about Indy's coat (00:06:56) so if you want to Indie coat these days (00:06:58) there's an embarrassment of riches at (00:07:00) your disposal you can go to westad (00:07:02) leather you can go to all sorts of that (00:07:04) made the originals and you can order one (00:07:07) from them uh in multiple different (00:07:09) Leathers um and they make it in some (00:07:12) really durable goat skin Cal skin lamb (00:07:14) skin I guess and the thing about most (00:07:18) produced coats from Raiders is that (00:07:20) they're way heavier than the actual coat (00:07:22) from Raiders which looks like a hearty (00:07:24) piece of leather it is so lightweight I (00:07:27) was just at Prop Store down in La they (00:07:30) had a Last Crusade jacket uh (00:07:33) and it's just so lightweight of all the (00:07:37) Indie jackets I have collected over the (00:07:39) years and I I've collected I think four (00:07:41) I've given away three I've kept only one (00:07:43) and the one I've kept is Todd's costumes (00:07:46) because Todd's costumes I thought was (00:07:48) the one that felt the closest to the (00:07:50) real run real one uh copen Cabana again (00:07:54) with all KS and extra letters uh what is (00:07:57) the biggest impact what has the biggest (00:07:59) impact in portraying emotion in a film (00:08:01) the actor actress ability the script the (00:08:04) score the (00:08:05) lighting so first of all I am not a (00:08:08) filmmaker I have not directed a film I (00:08:11) have not brought into being a thing that (00:08:14) is a film I have worked on many and many (00:08:17) of them so I don't know that I am the (00:08:19) person to answer this but um I did a I (00:08:23) did at one (00:08:24) point uh early on in my friendship with (00:08:27) Gad Del Toro I like when I knew I was (00:08:30) going to see him I would think up some (00:08:31) things that I really wanted to ask him (00:08:33) and one of the earliest ones I asked (00:08:36) was is there anything that you can see (00:08:40) while a movie is being made that you can (00:08:42) see that it's going to be great is there (00:08:44) any way to tell and he was like (00:08:46) absolutely (00:08:48) not you never know you can be working on (00:08:50) something you think is a piece of (00:08:51) garbage and it comes out and it's (00:08:52) incredible and the opposite is true but (00:08:55) he said for any film to be great there (00:08:59) is a commonality and it's that there's (00:09:02) at least one person who is that film's (00:09:05) champion and they have to bring it into (00:09:10) being if at least one person has it (00:09:14) wasn't that he said if a film has that (00:09:16) it's going to be great what he said was (00:09:18) every great film has at least one person (00:09:21) that championed it that took it on for (00:09:23) themselves to make sure that it (00:09:26) happened (00:09:28) um so it's a combination of all those (00:09:31) things it's a magic it's (00:09:33) a it's a Confluence right um and you (00:09:37) can't ever know that you've gotten it (00:09:39) all right until after you see it all (00:09:41) together I in a way so having just we (00:09:46) covered recently Lee enr's Incredible (00:09:49) tashion Book on The Shining it's over (00:09:51) 2500 pages in The several volumes that (00:09:54) make up this incredible object and I've (00:09:56) now read them all and one of the love (00:09:59) lest things that come out of it is um it (00:10:03) subverts The Narrative of Stanley kubri (00:10:05) as the genius who to totally knew (00:10:07) exactly what he wanted and went and (00:10:09) executed exactly that and then put it on (00:10:11) film for us to see nothing is further (00:10:13) from the truth reading all this material (00:10:15) about kubric what he really wanted what (00:10:17) he was really good at was knowing when a (00:10:19) story could create (00:10:21) drama and then he was really excellent (00:10:24) at setting up the circumstances for (00:10:27) putting actors into a situation where (00:10:29) could find the drama of a scene but he (00:10:32) didn't have a preconceived notion going (00:10:34) into a shoot it strikes me as to where (00:10:37) the drama was so he built a machine for (00:10:40) finding it and that involved bringing (00:10:42) actors in doing tons of takes with them (00:10:44) and then going into the edit room with (00:10:46) many different types of performances and (00:10:48) finding the film in (00:10:50) there I know that went far a field from (00:10:53) your original question but I do feel (00:10:56) like it it's gerine to that it is an (00:11:00) Alchemy but for any piece of art to be (00:11:04) really true it is always an Alchemy (00:11:07) there is no formula and there can be no (00:11:10) formula which is why going back to the (00:11:13) very first question about AI I'm maybe a (00:11:16) little less worried about AI (00:11:19) aesthetically than others I'm worried (00:11:21) about it from an employment standpoint (00:11:22) for all sorts of reasons that are (00:11:24) completely obvious but (00:11:28) aesthetically I feel like it's going to (00:11:30) be a near impossibility for an (00:11:32) artificially created intelligence to (00:11:35) show me a really distinct and particular (00:11:38) point of view that resonates with me (00:11:41) because they're not a person and I'm a (00:11:42) person and I respond to person's points (00:11:44) of view not machines at least up until (00:11:48) now um well I got (00:11:52) philosophical thank you so much for (00:11:54) watching if you'd like to support us (00:11:56) even further you can by becoming a (00:11:58) tested member uh details are of course (00:12:00) below But it includes all sorts of perks (00:12:02) and we're building them all the time you (00:12:04) get Advanced word and behind the scenes (00:12:06) photos of some of our projects questions (00:12:08) you get to ask direct questions during (00:12:10) my live streams and we have some members (00:12:12) only videos including the Adam Real Time (00:12:15) series of unbroken unedited shots of me (00:12:18) working here in the shop they are (00:12:20) weirdly meditative thank you guys so (00:12:22) much I'll see you on the next one