Revolutionary AI Image Control
AI Democratisation in Action, ControlNet, Stance Against ChatGPT + More!
Hello, my machine learning mavens and deep learning devotees! 👋
Deiniol here - founder of AI Suite.
I’ve spoken to a lot of AI artists recently.
Expecting them to be like me (slightly nerdy, doesn’t know how to hold a paintbrush properly, started and failed several startups before finding something that works), I was wrong. They come from all walks of life:
Paediatrics
Biology
Teaching
Horticulture
Data Analysis
These are people who have created tens of thousands, if not hundreds of thousands of AI-generated images over the last 9-12 months.
I’ve also talked to a number of my non-tech friends and family members about ChatGPT, and it's surprising how diverse their professions are. Among those that have used it are a carpenter, an electrical engineer, a retiree and a teacher. Even my humble gardener had at least heard of it!
This is AI democratisation in action. People are following their creative interests and passions now because they’re able to. So, perhaps they are a bit like me - happy that I now have several artworks around my house that I created. We are now doing more of what we love.
Anyways, that was too long for a Tweet so I had to put it here. On to the show (which is adapting based on feedback).
“The first AGI will be just a point along the continuum of intelligence”
📣 Highlights
JP Morgan Restricts Users from using ChatGPT - Bloomberg - No Paywall
Section 230 | The Supreme Court could be about to decide the legal fate of AI search - There is a possibility that AI-generated content would not be eligible for protection under Section 230.
OpenAI has formed a global services alliance with Bain & Company to help enterprise clients realise the value of generative AI.
⚙️ Tool 1
ControlNet | Stable Diffusion
Where to start!
ControlNet is a new incredible extension for Stable Diffusion. You can finally take control of image composition through:
sketches/scribbles
outlines
depth maps
and human poses
Essentially adding additional levels of control to Stable Diffusion images and, consequently, video/animation composition. It is a game changer and has pushed the boundaries of AI image and video creation even further, allowing for more artistic control.
As you can see from the images above, ControlNet is exceptional at:
Turning scribbles/drawings into images
Creating text out of spaghetti (or any other substance you can imagine)
Generating lifelike images of your favourite Playstation/Xbox characters
Colourising black and white photographs
Mimicking poses based on prompts & more!
This is truly an incredible extension for Stable Diffusion and, when combined with other tools such as Kaiber.ai and D-iD, a very reasonable amount of control and direction for AI videos is becoming easier and easier.
🔗 Important Links:
Auto1111 extension by Mikubill
For those who don’t want to install the extension but still want to play - click here.
Complete Guide for Posing Tools - courtesy of Olivio Sarikas
⚙️ Tool 2
Perplexity (Chrome Extension)
Perplexity have just updated their Chrome Extension and boy is it a doozy.
Ask questions about your current page, domain or the entire net
Share your answers with a permalink
Click to ask follow-up questions (albeit via the webapp).
Quickly ask questions from their toolbar and get answers with summarised sources and citations. It enhances the search engine experience by providing answers to users one click away. Because it answers complex questions by scouring multiple sources, like any LLM integrated into search, there is the possibility of wrong answers (pretty good on my end so far!).
No account is needed, no limit restrictions and no waiting for availability.
🧑🏫 Demo/Tutorial
Buildspace | Turn Yourself Into A Rapper Using AI
I like to consider myself a good rapper - now that I have ChatGPT and ElevenLabs…
Generate lyrics in the style of any rapper on any topic - all thanks to the powerful combination of GPT-3 and ElevenLabs’ API. Record your voice saying fancy words which are then transformed into high-quality voice-overs using ElevenLabs' advanced tech.
Don’t aspire to be the next Ice Cube? Imagine making Steve Jobs sing Hakuna Matata or generating your own stand-up comedy routine with ease. Or creating audio summaries of books or articles in your own voice. Pretty suite.
Build your own app today and unleash your inner Notorious Y.O.U.
Until next week, you deep learning devotees, code out!
Deiniol from Sydney (not Bing) 🇦🇺
Ps. I’ve been asked to provide more new tools so, ask and you shall receive! I’m toying with this new format, 3 highlights, 2 tools, and 1 demo/tutorial.
Let me know if you like it:
🏆 Sponsor Us
Would you be interested in sponsoring our next newsletter, which will be distributed to a diverse audience of AI enthusiasts, startup founders, aspiring entrepreneurs and artificial intelligence students?
Reply to this email and I’ll get back to you asap.