[This is a long read, I was also interviewed on this essay on the exceptional Macrovoices podcast recently if you’d like to listen:
https://podcasts.apple.com/us/podcast/macro-voices/id1079172742?i=1000625553151
Feedback on my interview:
“One of the most important podcasts I have heard.” @SteveBigpond
“(Macrovoices) This was one of your best podcasts. A strong contender for #1 actually. Thank you.” @kdogni]
On 22nd of August 2022, Skynet went online and started learning at a geometric rate.
At least, that’s what I’m sure it felt like for graphic illustrators.
On that day, Stable Diffusion, a deep learning text-to-image model was released. Like many others, I downloaded and started playing with it.
You’d type in a sentence like “man riding a motorbike, being chased by a bear”. What came out, at least for me, looked more like a nightmare- artistic yet often horrific images of people with extra body parts and warped faces in dream-like scenes. Reminiscent of a Picasso if drawn by Salvador Dali, the fusion of people and objects was striking, yet unnerving.
There was however something truly breathtaking about the software’s uncanny ability to manifest any typed-in concept into an illustration.