Bing Image Creator
The Six Five team discusses Bing’s Image Creator.
If you are interested in watching the full episode you can check it out here.
Disclaimer: The Six Five Webcast is for information and entertainment purposes only. Over the course of this webcast, we may talk about companies that are publicly traded and we may even reference that fact and their equity share price, but please do not take anything that we say as a recommendation about what you should do with your investment dollars. We are not investment advisors and we ask that you do not treat us as such.
Daniel Newman: So we’ve heard a ton about all this generative AI. We’ve heard about obviously Search, but that’s just one part of where generative AI is super powerful. You’ve probably heard about Dolly, probably heard about some of these different tool sets that can be used for everything from video creation to using Copilot for creating PowerPoints and more graphical. But we’re going to continue to see the growth of graphical and image creation. Pat at MWC, you and I played with Stable Diffusion, which was running on an Edge device with no connectivity showing a billion parameter AI model.
And what were we doing? We were doing image creation. Pat was drawing, what were you drawing? Something in a field or a leprechaun in a, I don’t remember. There’s a tweet out there. Pat will link you to it. But anyways, AI powered visual stories, knowledge cards, social tiles, graphics, can you do this kind of work in Bing or in Search or in Edge or in the browser? And the bottom line is yeah, Bing is now offering Image Creator. This week by the way, Adobe launched something called Firefly, which is its generative AI-
Patrick Moorhead: Cannot wait to use that.
Daniel Newman: All the stock and licensed images in the Adobe portfolio. And now Bing’s doing the same thing with create an image as prompt. Like I said, the story itself about Bing in AI and image and prompt is in my opinion more of a macro story of, you’re going to see this in Google soon. Bard is going to do an image creator, the Adobe one – and like I said, I’m only kind of pivoting here because I think there is a bigger story – is I like the ethical approach of Adobe only using its own and licensed data. Now Microsoft talks about its responsible and AI principles and AI standards in the press release to make sure that AI is being developed and deployed. Image creator is going to create a challenge though because I still think there’s a bit of a wild west in terms of where it’s getting its data from and how it’s creating images and how it’s creating attribution for those images.
And if it uses 500 different licensed or a creator images to create a new image, how does attribution get go get given to all of those others that have created and stamped and marked, or not marked, their work. But what I will say Pat, is I’m extremely excited because I don’t know about you, but I’m not a patient person. Meaning when I have to work with designers, nothing ever happens fast enough. And two is really the simplicity is I don’t tend to be all that particular, meaning I just want a high quality social tile or a high quality graphical card that looks professional. I’m really jazzed. I remember Canva, kind of the launch of Canva and something like that where this is 2.3.0 Canva now, where it’s like just tell it what you want. And now I need a food service that works the same way it’s called a restaurant, but I just need something a little bit more efficient. It’s like get the food real time, Jetson like. But yeah Pat – hey Microsoft just literally can’t help itself right now but just continue to launch stuff and it’s a pretty exciting time.
Patrick Moorhead: So I think this shows just how long Microsoft has been working on this and I’ve used the tool myself. By the way, related to the AI principles and responsible AI, this does abide with Microsoft’s responsible AI principles and responsible AI standard just like Chat did. I found it very conservative, but hey, we usually don’t do demos but I thought hey, why not just show what this does and how to do it. So as you can see here, you literally just type in what you want to see up here and then you press create. Over here are the amount of credits you have for boosting what you get. And it says create images more quickly with Boost. If you run out, image generation may take longer. What comes up here, Daniel? Monetization. Anyways, let me show a couple that I did that might some-
Daniel Newman: I want a picture of the host of the Six Five.
Patrick Moorhead: Yeah, so hunter jumper horse in heaven, photorealistic. Okay, how about a horse show? I thought that would be cool. Here is a sprinter in the four by 100 race diving for the finish line side view, digital art, make it more. And Daniel, you might enjoy this one. This is 2022 911 GT2RS jumping the Grand Canyon on a sunny day in Crayon. Thought that was cool. And here’s the photorealistic version of that. Now when it comes to people and animals, it’s kind of iffy, okay? 10 different colored French bulldogs. It kind of got it right here. But here’s where… Well let me show you the ones that I got. I thought it that were just cute. Check this out. Isn’t that cute?
Daniel Newman: Cuter to you than me. I’m not a Frenchy. I don’t-
Patrick Moorhead: Here is where it had some trouble and this is where it takes some hits, French bulldog riding a horse and jumping over a show jump it turned the horse into a dog and then put something on its back riding, which is kind of creepy. Here’s another creepy one. So yeah, I recommend it has some trouble with animals and it has some trouble with people. But anyways, I just thought that all of you-
Daniel Newman: How about the hosts of the Six Five Podcast?
Patrick Moorhead: Doesn’t do names buddy. And that’s where it’s conservative. I wanted Abraham Lincoln in a spacesuit and a spaceship and it won’t let you do any people. So very conservative and quite frankly, not nearly as fun as it could again. But-
Daniel Newman: We’re getting there. We’re getting there.
Patrick Moorhead: Yeah. No, no, there we are. So I’ve found my experience so far better on non-human, non-animal images. When you use the tags on the end like photorealistic or digital art didn’t see big difference. And there’s some hidden types that I found like watercolor and crayon. Can’t do fun ones with us, which would be fun. Again, Abraham Lincoln wearing a spacesuit and a spaceship digital art, I got a warning and they threatened to ban me if I do something else. I can totally see integrating this into PowerPoint. Final comments. And these are just some fact-based, this is based on Dolly model that comes out of the open AI. So great stuff. We saw LLMs with Bing Chat and now we have Dolly with Images and-
Daniel Newman: We got to do a side by side with Firefly, check out-
Patrick Moorhead: It would be fun, and also side by side on cost. Again, nothing wrong with price, but I saw some – I mean this is not a Firefly replacement or anything like that. This reminds me right now very “consumery” test tool. And this will be integrated into PowerPoint, which I’m going to fricking eat that thing up.
Daniel Newman is the Chief Analyst of Futurum Research and the CEO of The Futurum Group. Living his life at the intersection of people and technology, Daniel works with the world’s largest technology brands exploring Digital Transformation and how it is influencing the enterprise. Read Full Bio