I worked for 13 years on this problem (unlike the article, I focused on making films using the same skills used today in live-action, vs. having gamers start doing it), but in the end, couldn't actually get funded. (This was back in 2011, and we were raising $32 million, just over half of which was for hardware.)
I'm actually a filmmaker first, and learned programming after attending Siggraph in 1998 while going to film school and having a vision then of how to take Pixar's non-linear, iterative story/filmmaking process and use it to do live-action filmmaking. I knew it'd be at least 10 years before the hardware was ready, but there was a ton to do even before that.
Avatar got close to doing things the right way, on the way to becoming the highest grossing post-TV film, but was unable to iterate (due to lack of tech, but mostly because no one there actually knew that it'd be a good idea). Cameron got his first rough cut of Avatar more than two years into the process (!). By comparison, Pixar turns a new cut of their animated films around in a few months, and my process did so in about six weeks.
Anyway, all of this is to say: improvements in this area are going to have to happen the way the author is describing, even though the tech to actually re-build Avatar for around 1/10 of the cost has existed for three years now, because there is just no way to get the final solution funded directly.
Looking back, it should have been a warning sign to me that Pixar was funded out of Steve Job's own pocket to the tune of $50 million. Steve. Fucking. Jobs.
If he's the only guy willing to put money into this stuff (and to my knowledge, Pixar is the only company he put his own money into—it's also what made him a billionaire, BTW), then there's just no chance that Sand Hill, or worse, Hollywood, is going to fund it. It's not that they didn't like the idea or the tech, it's that a tech-based film studio is in a no-mans land as far as funding. It's never going to happen.
So I gave up, and now I'm CTO at a company doing a social networking app, and I just spent the last 36 hours prepping to finish up the custom streaming database I wrote from scratch which, according to my benchmarks, should do about 1 million writes per second AND 10 million reads per second—on a single box, fully encrypted between the client and server. Yee haw.
I can do that stuff almost in my sleep these days because the performance requirements of interactive, live-action style filmmaking where you can actually do physically-accurate lighting requires extreme optimizations in the software and hardware. That $16 million was a cluster than acted like a single machine, rendering inside a scene about 1.5 Petabytes in size, 24 frame a second. Social networking even at Facebook scale is a walk in the park compared to that.
But I get sad every time I read an article about this stuff, and remember what could have been. What a waste of talent to be doing this shit, when I could be making films. It reminds me of that guy who tweeted that all the best people in tech were spending their time trying to figure out how to make people click on ads. Yup, pretty much. Art? Who needs that?
But whatever, my plan now is to keep CTOing startups until I've got fuck you money and can just pay for a Pixar-like studio myself, out of my own pocket. Just like Steve Jobs did.
I'm actually a filmmaker first, and learned programming after attending Siggraph in 1998 while going to film school and having a vision then of how to take Pixar's non-linear, iterative story/filmmaking process and use it to do live-action filmmaking. I knew it'd be at least 10 years before the hardware was ready, but there was a ton to do even before that.
Avatar got close to doing things the right way, on the way to becoming the highest grossing post-TV film, but was unable to iterate (due to lack of tech, but mostly because no one there actually knew that it'd be a good idea). Cameron got his first rough cut of Avatar more than two years into the process (!). By comparison, Pixar turns a new cut of their animated films around in a few months, and my process did so in about six weeks.
Anyway, all of this is to say: improvements in this area are going to have to happen the way the author is describing, even though the tech to actually re-build Avatar for around 1/10 of the cost has existed for three years now, because there is just no way to get the final solution funded directly.
Looking back, it should have been a warning sign to me that Pixar was funded out of Steve Job's own pocket to the tune of $50 million. Steve. Fucking. Jobs.
If he's the only guy willing to put money into this stuff (and to my knowledge, Pixar is the only company he put his own money into—it's also what made him a billionaire, BTW), then there's just no chance that Sand Hill, or worse, Hollywood, is going to fund it. It's not that they didn't like the idea or the tech, it's that a tech-based film studio is in a no-mans land as far as funding. It's never going to happen.
So I gave up, and now I'm CTO at a company doing a social networking app, and I just spent the last 36 hours prepping to finish up the custom streaming database I wrote from scratch which, according to my benchmarks, should do about 1 million writes per second AND 10 million reads per second—on a single box, fully encrypted between the client and server. Yee haw.
I can do that stuff almost in my sleep these days because the performance requirements of interactive, live-action style filmmaking where you can actually do physically-accurate lighting requires extreme optimizations in the software and hardware. That $16 million was a cluster than acted like a single machine, rendering inside a scene about 1.5 Petabytes in size, 24 frame a second. Social networking even at Facebook scale is a walk in the park compared to that.
But I get sad every time I read an article about this stuff, and remember what could have been. What a waste of talent to be doing this shit, when I could be making films. It reminds me of that guy who tweeted that all the best people in tech were spending their time trying to figure out how to make people click on ads. Yup, pretty much. Art? Who needs that?
But whatever, my plan now is to keep CTOing startups until I've got fuck you money and can just pay for a Pixar-like studio myself, out of my own pocket. Just like Steve Jobs did.