Prompt on Dream Machine: Astronaut dancing in zero gravity holding a gay flag.
It would be a fun art project to train a set of prompts on nothing but porn and funny cat videos.
1 Like
I mean, the possibilities of âcat got your tongue?â alone seem intriguing.
Yeah our gig is over. What could I add to that, and for my rate?
Volvo - for artificial lifeâŚ
5 Likes
Please tell me youâre joking.
Assuming this is the origin (seems plausible) - see the description:
From Reddit
Idea, prompting, post-production by me.
This is a test for trying out the RunwayML Gen-3 Alpha model for commercial purpouses.
All the shots are generated with RunwayML using text-to-video, some small details were retouched in AE after the edit was done (logo, license plate, or if I really liked a shot but there were two cars present, etc).Altogether I spent less than 24 hours working on this video - Gen3 is only out since 1.5 day ago - this includes generating the clips, selecting them, editing them to a coherent story, retouching and sound design.
Is this a big step-up from Gen-2? Definitely! Gen-2 was a great experiment but Gen-3 is something that will change the industry. Pricing wise some people say itâs expensive but I think if you go for the unlimited plan itâs worth the money.
I just canât wait for the feature updates including image prompting and director mode!
If AI is good at something, itâs creating interesting back plates where detail and interaction is not as critical. I would assume in a real-world scenario Volvo would actually still film the car, the interior, and some other critical detail and then comp the two together. I find it hard to image that they would entrust such critical product details to AI. Plus AI is trained on yesterdayâs models, not cars that arenât yet on the market. Of course they could augment that training, but that would require videos of it (a classic chicken and the egg).
I think thereâs also a place for pre-viz.
But it will still take talented artists to bring this over the finish line.
Yes, and only kinda.
2 Likes
Lots of commercials for this yearâs car are shot with last yearâs model.
and then they pay to update the trim/headlights/mirrors/tires/dashboard/etc. in postâŚ
yep
will machines be better at it - probably, if having seven mirrors per door and hairy feet is acceptableâŚ
will machines get better at it - oh yes - have no doubt
for sure.
I guess what annoys me about this video is A) Iâm a little jealous and B) the artist is hiding the ball, so nothing specifically about the AI part of it.
4 Likes
that video is from 7 years ago. They may have had an opportunity to be leaders. Now, I havenât seen a single AI innovation come from Japan. In 5 years, there will be 1000s of Studio Ghibli style movies coming from little Johnny and sweet Jane pressing buttons on their iPhones.
1 Like
They wonât be Miyazaki films though. Aside from that, I think sweet Jane and little Johnny are an interesting case. You should look at reading trends in gen z right now. Not what theyâre reading but how. I think there are going to be more and more people looking for ways to get away from phones in the coming years. I actually imagine a whole movement of smart phone natives saying âthese things poisoned us, they made us depressed and anxious and we need to cut the cord.â Weâll see. This is from a ridiculous article in the guardian about reading being âsexyâ to gen z, but I was struck by the number of physical books and the push for access to that. I think the idea of filling the slop trough for media and entertainment isnât a winning strategy. Besides, thereâs already a never ending slop trough and they call it TikTok, people donât seem to be clamoring for anything new when it comes to addictive content.
3 Likes
Interesting take from an investor (via Fortune - may hit paywall):
AI is effectively âuselessââand itâs created a âfake it till you make itâ bubble that could end in disaster, veteran market watcher warns
Thereâs no avoiding the hype surrounding AI these days. Promises of new developments like personal robot assistants and miracle cancer cures are ubiquitous as executives take every opportunity to emphasize their AI chops to enthusiastic investorsâand slightly less enthusiastic consumers.
Not everyone has been blown away by the AI fanfare, however. James Ferguson, founding partner of the UK-based macroeconomic research firm MacroStrategy Partnership, fears investorsâ AI exuberance has created a concentrated market bubble thatâs reminiscent of the dot-com era.
âThese historically end badly,â Ferguson told Bloombergâs Merryn Somerset Webb in the latest episode of the Merryn Talks Money podcast. âSo anyone whoâs sort of a bit long in the tooth and has seen this sort of thing before is tempted to believe itâll end badly.â
The veteran analyst argued that hallucinationsâlarge language modelsâ (LLMs) tendency to invent facts, sources, and moreâmay prove a more intractable problem than initially anticipated, leading AI to have far fewer viable applications.
âAI still remains, I would argue, completely unproven. And fake it till you make it may work in Silicon Valley, but for the rest of us, I think once bitten twice shy may be more appropriate for AI,â he said. âIf AI cannot be trustedâŚthen AI is effectively, in my mind, useless.â
Ferguson also noted AI may end up being too âenergy hungryâ to be a cost effective tool for many businesses. To his point, a recent study from the Amsterdam School of Business and Economics found that AI applications alone could use as much power as the Netherlands by 2027.
I tend to agree mostly. There is a lot of hype and crazy dreams. Some good will eventually come from it (as did from the players in the .com bubble). So itâs not all for naught. But a healthy grain of salt is warranted.
For me I think the problem is actually going to be more the economics rather than the question whether eventually some model will make more realistic videos. Once we actually have to pay for all this at full price (without the investor discounts), we might find out that what we used to do actually provide more predictable results for the same or less.
You already saw some stories that the AI industry is trying to lock up all the available nuclear power in the US so they can get enough juice without getting bashed for adding to climate change at the worst possible moment in that crisis.
2 Likes
I agree with what youâre saying for sure in terms of the economic side. I also though am skeptical that the uncanny valley can just be leapt over very easily. I think these prognostications of âjust wait, in 6 months, 2 years, 5 years,â etc. seem to just assume some fairly linear progression. And I donât know why weâre so certain of that. When people say, âremember, this is the worst it will ever be,â I always tend to think, ok, if I need to travel six inches, moving one inch is a huge step forward, but if I need to go 1000 miles, one inch is barely movement.
Yes, I donât think many folks saw the leap LLMs brought with them coming. AI is not a fresh technology, but it just made a huge leap in 2022. Some of us had seen previews elsewhere (as in the Neeva search engine), and then ChatGPT broke it wide open. But how many years was that in the works? Decades. They talked about neural networks when I was in college in the early 90s. Itâs only that they suddenly were able to scale them differently - and came up with the right arrangement of the nodes.
But just as you said - this is far from linear, and I cringe everytime I see someone write âoh this will be solved shortlyâ and they fully mean that, because of how quickly new AI tools hit the market, and they donât have line of sight of what happens behind the curtain.
There could be another breakthrough in 6 months and those problems could all be fixed. Or it could be another 10 years. Nobody knows. However, I donât think the fixes are easy by any means. With the massive amount of resources and focus on it, we might have seen more progress than we did if 6 months is realistic.
3 Likes