In the video towards the bottom of the page, there are two birds (blue jays), but in the background there are two identical buildings (which look a lot like the CN Tower). CN Tower is the main landmark of Toronto, whose baseball team happens to be the Blue Jays. It's located near the main sportsball stadium downtown. I vaguely understand how text-to-image works, and so it makes sense that the vector space for "blue j…
Have you seen fal.ai/dynamic where you can perform image to image synthesis (basically editing an existing image with the help of diffusion process) using LCMs to provide a real time UI?
I'm still puzzled as to how these "non-commercial" model licenses are supposed to be enforceable. Software licenses govern the redistribution of the software , not products produced with it. An image isn't GPL'd because it was produced with GIMP.
The license is a contract that allows you to use the software provided you fulfill some conditions. If you do not fulfill the conditions, you have no right to a copy of the software and can be sued. This enforcement mechanism is the same whether the conditions are that you include source code with copies you redistribute, or that you may only use it for evil, or that you must pay a monthly fee. Of course this enforce…
It also somewhat depends on open legal questions like whether models are copyrightable and, if so, whether model outputs are derivative works of the model. Suppose that models are not copyrightable, due to their not being the product of human creativity (this is debatable). Then the creator can still require people to agree to contractual terms before downloading the model from them, presumably including the usage limitations as well as an agreement not to redistribute the model to anyone else who does not also agree. Agreement can happen explicitly by pressing a button, or potentially implicitly just by downloading the model from them, if the terms are clearly disclosed beforehand. But if someone decides on their own (not induced by you in any way) to violate the contract by uploading it somewhere else, and you passively download it from there, then you may be in the clear.
In the video towards the bottom of the page, there are two birds (blue jays), but in the background there are two identical buildings (which look a lot like the CN Tower). CN Tower is the main landmark of Toronto, whose baseball team happens to be the Blue Jays. It's located near the main sportsball stadium downtown. I vaguely understand how text-to-image works, and so it makes sense that the vector space for "blue j…
> Has anyone come across a solution where model can iterate (eg, with prompts like "move the bicycle to the left side of the photo")? It feels like we're close.
Emu can do that.
The bluejay/toronto thing may be addressable later (I suspect via more detailed annotations a la dalle3) - these current video models are highly focused on figuring out temporal coherence
A seemingly off topic question, but with enough compute and optimization, could you eventually simulate “reality”? Like, at this point, what are the technical counters to the assertion that our world is a simulation?
The brain does simulate reality in the sense that what you experience isn't direct sensory input, but more like a dream being generated to predict what it thinks is happening based on conflicting and imperfect sensory input.
Example vision: comes in from the optic nerve warped and upside down and as small patches of high resolution captured by the eyes zigzagging across the visual field (saccades), all of which is assembled and integrated into a coherent field of vision by our trusty old grey blob.
It's really not. Don't get me wrong, this is insanely cool, but it's still a long way from good enough to be truly disruptive.
One year. All of Hollywood falls.
No offense, but this is absolutely delusional.
As long as people can "clock" content generated from these models, it will be treated by consumers as low-effort drivel, no matter how much actual artistic effort goes in the exercise. Only once these systems push through the threshold of being indistinguishable from artistry will all hell break loose, and we are still very far from that.
Paint-by-numbers low-effort market-driven stuff will take a hit for sure, but that's only a portion of the market, and frankly not one I'm going to be missing.
Porn will be one of the main use cases for this technology. Porn sites pioneered video streaming technologies back in the day, and drove a lot of the innovation there.
Looks like I'm still good for my bet with some friends that before 2028 a team of 5-10 people will create a blockbuster style movie that today costs 100+ million USD on a shoestring budget and we won't be able to tell.
The first full-length AI generated movie will be an important milestone for sure, and will probably become a "required watch" for future AI history classes. I wonder what the Rotten Tomatoes page will look like.
"I wonder what the Rotten Tomatoes page will look like"
Surely it will be written using machine vision and llms !
A seemingly off topic question, but with enough compute and optimization, could you eventually simulate “reality”? Like, at this point, what are the technical counters to the assertion that our world is a simulation?
> Like, at this point, what are the technical counters to the assertion that our world is a simulation? How about this theory is neither verifiable nor falsifiable.
The general concept is not falsifiable, but many variations might be, or their inverse might be. E.g. the theory that we are not in a simulation would in general be falsifiable by finding an "escape" from a simulation and so showing we are in one (but not finding an escape of course tells us nothing).
It's not a very useful endeavour to worry about, but it can be fun to speculate about what might give rise to testable hypotheses and what that might tell us about the world.
In the video towards the bottom of the page, there are two birds (blue jays), but in the background there are two identical buildings (which look a lot like the CN Tower). CN Tower is the main landmark of Toronto, whose baseball team happens to be the Blue Jays. It's located near the main sportsball stadium downtown. I vaguely understand how text-to-image works, and so it makes sense that the vector space for "blue j…
I also wonder if the model takes capitalization into account. Capitalized "Blue Jays" seems more likely to reference the sports team; the birds would be lowercase.