Artificial Intelligence, Technology & Entrepreneurship
7 predictions for the Jasper AI launch event
AI evolves tomorrow. Here’s what you need to know
Jasper is a cutting-edge company about to change the future of AI again. The Jasper AI team has been working hard on their next big release and they’re finally ready to show the world what they’ve been up to.
The virtual launch event is scheduled for October 18th 2022 @ 1:00PM CDT, and expectations are high. If you want to attend, you should register as soon as possible — seats are limited and are expected to fill up fast!
[UPDATE: Here’s my recap of the Webinar]:
Here are 7 predictions for what we think the Big Reveal might be:
1. We’ll finally see commercial Subject-Driven Generation in Text-to-Image AI art
This would mean that your images could have a consistent subject across AI-generated art, like a specific person or object. While I posted a Proof of Concept for using Jasper in this way, this has so far only been perfected in labs, like Google research’s DreamBooth. If commercial SDG comes to Jasper, it would be a game-changer for the platform.
2. We’ll get typographical imagery in AI art
One thing AI-generated imagery struggles with is typography, because it works on visual data rather than comprehension of written language. We think the Jasper team might finally have a solution. If Jasper has cracked this they could open up a whole world of possibilities for AI-generated art, like creating images with perfectly typeset quotes, logos, or product labels.
3. Video generation/fully animated AI images
The Brain Team at Google Research has been working on a project called Imagen. Imagen is an AI model that generates interleaved spatial and temporal video super-resolution models. Basically, it makes a video from a simple text prompt (“a dog riding a skateboard in Times Square”). I think that Jasper AI may finally have a solution that works well enough for commercial use. This could allow for really interesting applications, especially if you could generate a video clip from an image input.
4. Improved human pose estimation — including photorealistic hands!
This is something that’s been happening in the research world for a while now, but we think it’s finally ready for prime time. Accurate human pose estimation is notoriously difficult to perfect. AI particularly struggles with occluded human poses in naturalistic settings (i.e. legs under tables and chairs). But if Jasper can pull it off it would be really amazing. This could be really useful for things like augmented and virtual reality applications, and for more realistic 3D character models. Speaking of 3D AI images…
5. We’ll see the first Jasper 3D AI-generated images
3D diffusion models are going to be essential for things like augmented reality applications or generating more realistic 3D character models. 3DiM is leading the way in this field, for medical and human-robotic interaction.
Jasper AI could give us our first taste of what 3D AI images will look like for regular consumers. Again, I made a Proof of Concept to show that Jasper Art’s underlying model is capable of comprehending 3D space, but it’s not quite ready for public consumption yet.
If Jasper AI has mastered this, it would be a HUGE breakthrough.
6. The Jasper Voice system will be unveiled
I almost put Jasper Music here, but honestly, that’s too basic. AI-generated music is already offered by Jukebox — OpenAI, SOUNDRAW, and AIVA.
Instead, I’m predicting natural human voice emulators. This has many applications, from audiobooks and podcasts, to voice-over work, to translations. I also predict it will be able to adopt the tone, intonation, and mannerisms of a specific individual, which has all sorts of implications.
On the positive side, it could be used to create realistic characters for video games or movies. It could also be used to generate accurate translations of speeches or other spoken documents. On the negative side, it could be used to create fake audios that could be used to spread disinformation or discredit someone. Overall, I think the potential applications of this technology are fascinating.
7. Dave Rogenmoser, CEO of Jasper, will appear as a fully-animated 3D-generated hologram, voiced by AI
Hear me out. Holograms are not that far-fetched. Tupac was resurrected as a hologram and we’ve been able to generate 3D images of people from 2D photos for a while now. I think that if Jasper has mastered 3D generation, voice emulation, and video generation, they could use it to create an AI-driven hologram of their CEO that can be used for marketing events, media appearances, and even internal company presentations. Imagine if you could have a conversation with a hologram of Dave Rogenmoser!
So those are our predictions — what do you think? Do you agree? Let us know in the comments!
Who knows what the Jasper team has actually been working on? We’ll be sure to keep you updated on all the latest from the Jasper AI launch event!
We’ll find out soon enough! In the meantime, if you want to attend and be among the first 3000 people to know and get ahead of the game: Remember to register!
You can access your invitation below:
