D-ID's AI Magic: Transforming Still Photographs into Dynamic Videos

0:00So DID is a Tel Aviv based startup and they have their kind of recognized for their technology behind a bunch of different viral videos of animated family photos. But they essentially have expanded their platform to mobile users. After initially, I think they just launched this as like a web platform or whatever, but they have the DID's creative reality studio that now allows users to pretty much bring images, plain images to life with AI generated videos. And now you can do this on your phone. So that's even, I think that's even more accessible for a lot of people. But the company had initially kind of eyed corporate sectors for applications like training or internal and external communications as well as product marketing um i think with the new mobile like the new mobile app they have users can now really quickly access did services by you know you just download the app log in or whatever um you can opt for a pre-existing digital persona which did has created or you can just upload a photo from your device so it's the same way that Lenza allowed you to upload a handful of photos of yourself and then it could generate like cool AI images of you in all sorts of places.

1:11This is the same idea except now you just upload an image of yourself and from that image they can create like a talking head video of you saying or talking or doing which is really cool technology especially if you think about once people start integrating this with things like 11 labs it is doing like a voice clone or something if they do it or if they do their own version so um i think the the interface is fairly user-friendly on this um it lets you put text in 120 available languages they have both male female voice choices um and i don't think they do like voice cloning yet so i'll be curious to see if they go with 11 labs to do that because it'd be really cool if you could like i create like a voice clone on 11 labs of yourself and then go here and get it to be like upload an image and have it be like a talking head video Now you have a talking head video that's you talking, and all you had to do is give 11 Labs a couple seconds of audio and give this a picture, and all of a sudden you're able to do this.

2:06I think users can also customize the voice tone, so you can make it really cheerful, friendly, terrified, whispering, like a bunch of different options. That's kind of cool. I was looking at some demos of the tool. It's honestly impressive. It's pretty cool. It's not perfect, right? I think the lips syncing to the audio is not perfect, but it's pretty close. It's pretty solid. It's obviously going to get better. So I'd say just like imagine it being perfect because it will be perfect soon. You know, I think in a conversation about some of the underlying technology, their CEO, Jill Perry said, quote, DID integrates a blend of unique and open source AI techniques central to it is a model adapt at crafting video frames guided by audio signals or products powered by a potential API can render videos at a staggering 100 frames per second, which is could triple the speed of typical real time rendering, which honestly is pretty impressive.

3:04According to Perry, their CEO, while you know ai enthusiasts are the primary target right now the broader appeal um is quote it's a tool for those desiring videos of digital personas be it their digital twin historic figures fictional characters or even ancestors what the app truly offers is an affordable scalable video production solution that's accessible everywhere and like when he said the whole ancestors thing all of a sudden that made me think of like there was like a thing a while ago where like you uploaded a picture of like one of your ancestors and it kind of gave it like uh like a motion like the face was kind of moving and you could like imagine what they looked like a little bit better imagine this like you go find your great great grandfather's journal or great great grandmother's journal upload all the text in of like the stuff they said and then you get a bunch of pictures of them that are like talking head pictures and like you could probably just create an ai version of like you want one of your ancestors and chat with them and ask them questions about their life I know my wife, her grandmother who passed away when she was like very little, she didn't really know her, but she kept extensive journals throughout her whole life, you know, documenting everything.

4:15And it would be fascinating, I'm sure, for my wife to be able to upload like the 20 journals that this lady kept and just be able to talk to her, ask her anything about her life and her experiences for advice. It would be so fascinating, and I guarantee that it will get there. This is definitely an area that this technology will be used for. I think specifically the company DID emphasized some of the ethical uses of its technologies. So video can span up to 10 minutes, but the company is trying to ensure their platform isn't exploited or creating deepfakes, that kind of stuff. So they've made a bunch of commitments to honor copyrights, adhere to some different AI guidelines, and refrain from licensing the technology to sectors that are controversial.

5:04They said they're not going to license it to political entities, adult content publishers, and weapons manufacturers, which is kind of interesting and, I mean, understandable for a lot of that. So before its mobile introduction, the desktop variant of Creative Reality Studio had around 60 ,000 registered new users daily, and they were creating about 150 million. I think they made over 150 million videos. So the services API has around 200 ,000 users, which is really interesting that the users from the API is a lot bigger than just the desktop app. The mobile iteration is set to, okay, one thing I'm going to say is I'm not 100 % clear when they say that.

5:54Like they said the API has 200 ,000 users. I don't think that's 200 ,000 companies that are using their API to like produce content or whatever. I think that it's like they gave out an API, a handful of companies have implemented it, and those companies have gained 200 ,000 users. So yeah, I could be wrong on that, but that's my assumption there. Open to corrections, do your own research. In any case, I think this whole mobile side of it is definitely going to be a lot bigger of a net catering to the whole mobile centric demographic. I think the move is essentially aimed at current users seeking to kind of streamline the mobile experience and potentially capture a bigger audience.

6:36did's mobile services i think will operate as a subscription model so it's going to be about six bucks a month um and then existing subscribers of the desktop platform can like really easily continue their plan um and for those i think they get like a 14 day free trial period which is you know cool um i think despite being uh headquartered in tel aviv did's like very global they have a staff that like pretty much their staff's all around the world um and they said they're doing this to ensure uninterrupted services even during geopolitical disturbances. They specifically highlight the Israel-Hamas conflict, which is interesting, sad, you know, that they have to do that.

7:18I'm sure there's a ton of startups in the Ukraine that I think have faced a lot of very similar things. So I'm sure they're seeing kind of that same roadmap and seeing what happened there and making sure to set up their company to be successful regardless. list um perry their ceo said quote our team both in israel and globally are committed to delivering unmatched support particularly our ukraine team has been exemplary in showcasing their solidarity with the circumstances in israel so last year i think after successfully raising 25 million dollars in a series b round did they made headlines um with its deep nostalgia project and its offshoot Livestory.

7:58So both ventures centered around animating historical family photos. So you probably have seen like these, they were, they're kind of quite viral at the time. Livestory incorporated audio. So this initiative saw a ton of engagement. A bunch of different people were using animated narratives of distant family members. DID's ongoing projects, I think, also include a photorealistic interface for chat GPT, which is, I think, kind of cementing its place in the whole, you know, in this space of AI driven innovation, this is going to be a really interesting company to follow in the future. They're doing some really phenomenal stuff.

8:34So I'm excited to test them out myself and follow them as they continue to evolve because right, they, they started as just like they would animate a picture and then it was animated picture. Now it can talk and it'll be interesting to see where that goes. It's going to be, you know, an animated, uh, actor that you're going to be able to put in your own movie and And they'll probably play a role in that whole audio-video kind of revolution we're seeing with Generative AI. So definitely an interesting company to follow.