This post is inspired by the fact that I do a lot of things, and I should share more of them. A periodic recap seems like just the ticket.
What are some of the things I did this week?
Explored new ai model styles.
Trained a few styles models on a couple of different platforms to compare.
Explored a wide range of visual ideas related to a wide variety of subjects including the latest upcoming holidays: St. Patricks day and Easter.
Added some new vector stock icons visualizing ai, machine learning, automation, technology in business related concepts to Creative Market. https://creativemarket.com/sjagiello/291114769-AI-API-Automation-Icon-Set-%E2%80%93-100
Added some Easter illustrations to Shutterstock.
Explored gpt’s ability to draw scenario specific outfits, such as the most appropriate outfits for Summer in the desert climate of Arizona. Useful to prep for the hot weather ahead. https://www.instagram.com/p/DHL_fBQP6ZR/?igsh=MTJiZnRhdW5rNGRnaw==
Added some things to Instagram such as:
Sora scene stitching tests to see how much time that takes vs output quality https://www.instagram.com/reel/DGkKFp-uwLK/?igsh=NjVxc2c4dHh3d2lr
Sora cartoon from seed image test to see if it would infer a specific cartoon motion animation style based on a specific illustrative image https://www.instagram.com/reel/DG5-lONAMgp/?igsh=MWE2YnM4ZmZ6M2Vjbw==,
My projects over last year have been heavily UI design related and not public facing:
Illustrator and Figma screens and prototypes, design token systems, visualizing the right content for the right product audience, responsive design, creating presentations, evaluating new design tools, created a lot of front-end code sketches for ui motion, comparing it to Lottie's, designing mobile-first thinking about accessibility requirements for healthcare web apps, strengthening a pov on which ui frameworks are best for healthcare compliance (Ibm Carbon is a favorite https://carbondesignsystem.com/), and why (sustainable accessibility in a responsive design system can be costly to create from zero/balance that out by leveraging lessons from the most compliant solutions).
And, to balance that out, I switch to a completely different genre, like illustration of cartoon style characters, holiday themes, or things unrelated to ui to keep it fresh.
It might look like:
Design responsive screens for part of a web app.
Then switch creative gears, yet still working in vector, and study the techniques used by one of my favorite artists. For example, Mary Blair. See if it’s possible to capture the same degree of style or mood in vector.
Rest one side of the brain and use the other side for a bit. It works. Clarity happens in background on difficult technical issues that maybe don’t have a specific right or wrong answer. Switching creative modes or subjects until there’s a natural break works well.
I also made a relatively quick birthday AR for someone. It doubled as an ar immersive user triggered animation test with proximity and without proximity. https://www.instagram.com/p/DGwKQWSyNlY/?igsh=MW42NDlydGM0cG84dg==
I was thinking about how it’s easy to invoke JavaScript libraries in html and wondered if there were a machine learning one I could invoke using codepen https://codepen.io/sjagiello on my phone, and if I could use hugging face diffusers.js. Then I started seeing if any image generators could be invoked, criteria passed, or embedded, etc. https://codepen.io/sjagiello/pen/XJWNvEB
Would be interesting to be able to pass one prompt to 20 different image generators at once and see the results on one page. Not really something to sketch on my phone or in codepen I realized, so it became more of a list. https://codepen.io/sjagiello/pen/emYgvBG
Also sketched a few ui motion transitions in answer to a question.
Sketched a geospatial earthquake 3d data vis because my mom is going to Greece and I wanted a personal quick page to monitor the earthquakes around the Santorini caldera. https://codepen.io/sjagiello/pen/KwKVBBR
Created a Design System Markmap https://codepen.io/sjagiello/pen/azbdxyd Basically the idea is that get any structured content into code and it can be instantly reused as variables in a structured system. So, markmap reads markup and displays the content. It’s a js lib so invoked it in code to get a quick sense of it. Would be useful to create an agent and ask it to step through the markmap and perform the specific task required to create that part of the system.
Tbc