"Generative AI" is mainstreaming seemingly everywhere since coming on the scene in 2018 as described by a paper from OpenAI.
Introduced in 2020, we've had GPT-3, a text-to-text auto-generation system, commonly used in natural language translation where context is integrated into the translation; and for text generation from simple prompts. (*)
The next thing up the generative AI ladder is text-to-image. And for that the AI folks have multiple competing systems: DALL-E-2, Midjourney, and Stable Diffusion. And some have started to 'generative AI' video built on top of Stable Diffusion.
Microsoft has had a strong investment position in generative AI for some years with its partnership with OpenAI. Now we see some of this effort mainstreaming for the casual user. Several posts like this one have announced the coming generative AI capability in MS Powerpoint -- built off DALL-E -- for generating clip art from text prompts.
In combination with text-to-text for presentation animation, this could be a biggie for presentations, proposals, marketing material, and really anywhere imagery is used throughout the project. This is especially interesting when you get to architecture and physical design.
Of course, for the casual user, this may be a bridge too far, meaning that without some experience the tweaking required to get a clip art that is really illustrative of your point, this may be too much trouble -- at first.
But for the power user, AI may be just the trick for really clever and appealing art to illustrate project materials.
--------------------
Like this blog? You'll like my books also! Buy them at any online book retailer!
