ICYMI: Chevy is feeding his original photos to an AI to create an NFT series (and documenting it all here). Follow on IG and Twitter for first looks.
Today I have a double dose of discovery and learning to fuel your AI/NFT degeneracy:
-
A 1000 ft view of Midjourney
-
A big ol’ discovery!
The 1000 ft view of Midjourney (aka how MJ works, so you can start tinkering with ease).
The free (but messy) option:
Get a Discord account → Join the Midjourney Discord channel → go to one of the “#newbie” bot channels in the left column → type ‘/imagine’ followed by your text prompt.
Only problem with this option is there’s ~13M members in the Midjourney Discord, with ~1.5M folks active at any one time.
Which means these channels are a mess of constant prompts and images being sent/received that quickly burry the results of your prompts.
Here’s how to calm the noise…
The free (and chill) option:
Refer at least 1 person to Web3 Daily (using this link: {{subscriber.rh_reflink}} ) → get an email from us inviting you to join our Discord → Go to the “Midjourney” channel in the left column → type ‘/imagine’ followed by your text prompt.
(It’s much calmer and easier to track/create this way).
Today’s big discovery!
You can send Midjourney multiple reference images.
Which means if you have a ‘reference image + text prompt’ combo that is serving up the goods, you can produce new-yet-familiar creations by throwing in additional reference images.
Seems like a small discovery – but it really lets you get on a roll once you find a ‘reference image + text prompt’ that works.
For example:
This image, along with the prompt ‘Nasa, Saturn V, Soyuz, nasa uniform, speaker system, satelite, radio tower, photo by Larry Sultan and Petra Collins’ was delivering some absolute bangers…
Like so:
