New Video: GPT-3: The End of Writer's Block (?) [GPT-X, DALL-E, and our Multimodal Future]
Can GPT-3, DALL-E, and multimodal AI models of the future pull you out of creative tight corners? How can multimodal AI help you with your project's motivation/psychology and long term success?
Watch today’s video to find out!
YouTube Transcript (SPOILER WARNING)
Do you ever wish you could take a step away and get a fresh look at your project as if it was the first time? Do you ever wonder how your audience will perceive your work or if they are able to follow along with its progression?
My friend James is already observing that at his Sci Fi writing startup, Sudowrite, he’s seeing an interesting phenomenon where the writers are using his GPT-3 powered AI tool to get feedback on their work and to determine if the underlying themes are actually making it through to a theoretical audience.
He relates this feedback technique to rubber ducking which is a technique programmers use to solve issues in their code. It’s basically where you talk out loud about your coding problem and often you end up answering your own problem with a solution.
When I think about it, I’ve already used GPT-3 in similar ways. I’ve used it in a lot of my YouTube scripts - to get feedback on my ideas, come up with new ones, or to help me restate my ideas in simpler ways. Not only that, many people in the GPT-3 community have used GPT-3 to take on a different role where it could teach, provide advice, summarize, and all kinds of creatively helpful perspectives and use cases beyond just content generation.
What I’m trying to say is that generating content is not the only role multimodal AI models can take up for us in the future. For example, by simply acting like a non judgmental audience member, multimodal AI could provide the audience feedback loop creatives have always needed, leading to better creative works.
Artists have a lot of different approaches for this exact problem to help them “get out of their own heads”. As they’re working, concept artists like to take their art work and flip the canvas horizontally to literally get a different perspective on their art. Copy editors like to take existing content and read it out loud but in backwards order to evaluate the same content in a new way. Personally, even I like to dramatically change the formatting of my writing, and just by doing that, it looks foreign enough to me that I can start to look at my piece differently as if it’s the first time.
Going back to the futuristic multimodal photo editor app concept, let’s say we have an existing piece going but we’re not sure how it will be perceived by a first-time audience. So, inside of the editor with an existing image ready, you can see there’s a realtime audience feedback section on the right, however, for our purposes, I am going to go to the analysis and feedback section at the top. You can see I have multiple options to choose from, where multimodal AI will analyze the work and give us feedback. If you’ve seen the composition video, you already know about compositional analysis. But many of these other options allow us to look at our work from different ways.
How does this image fit into the current pop culture zeitgeist? Is it coherent or original compared to anything else seen before? By clicking on any of these options, AI will give us a written critique on our work instantly. Let’s click Thematic Success to get some feedback. Awesome - It’s really that handy.
The Key Idea
AI can serve many roles for you creatively ranging from potential audience member, to focus group member, to coach, or even to creative collaborator. Learn to leverage multimodal AI by actively soliciting feedback and analysis for your art in real-time as you are making it.