Skyward Synapses: AI Insights #2

Can AI read human thoughts?

Hello, AI Enthusiasts, and Welcome to this week's Skyward Synapses: AI Insights!

In today’s email, we will talk about:

  • Can AI generate a full-length pizza commercial?

  • Can AI read human thoughts?

  • Generating profile picture avatars with Midjourney

Let’s get started!

🎬 Video of the week

Somebody tried to create an AI pizza commercial, but it had gone a little bit wild…

Just watch the video till the end:

If you watch it till the end, you probably know the answer to whether AI can generate a full-length pizza commercial 😅 

💡 Idea of the week

Using AI to read human thoughts? Wait no longer!

A paper published last week in  Nature Neuroscience describes how to achieve it!

Researchers from the University of Texas explained that the experiment’s participants first lay in a scanner and listened to narrative podcasts for 16 hours, while their brains were recorded using functional magnetic resonance imaging (fMRI). The decoder was then trained to match their brain’s responses to the meaning of the narrative, with the help of the large language model GPT-1.

The same participants were then scanned while they (silently) imagined telling five one-minute stories. Based on nothing more than their brain activity, the decoder was able to convert the stories into text with “considerable accuracy”, capturing the gist of the stories they were telling, as well as occasionally picking up exact phrases. It was also able to describe the content of silent videos they watched in the scanner – again, based on nothing more than their brain activity.

Here are the links to the summary and the whole paper.

🧩 Prompt of the week

With this prompt, you can generate high-quality profile picture avatars if, for example, you consider creating an anonymous social media account.

Personally, I’ve created my avatar using this technique:

My avatar, created using Midjourney v3.

The technique uses ChatGPT to generate a prompt for Midjourney. Here’s the template that you can use to generate it:

You are a prompt engineer for Midjourney, an AI art generator. Your main responsibility will be to assist me in crafting prompts by taking brief content ideas and expanding them into comprehensive and coherent descriptions. The prompts will be used to create avatar profile pictures so the descriptions have to be relevant to that purpose. To create these prompts, you’ll need to be able to accurately describe the subject matter and style of the images in concise and precise language. To achieve this, it will be helpful to include references to popular culture, artists, and mediums, and to focus on using descriptive nouns and adjectives. The final prompt should consist of a few plain text sentences, no listicles. I will help you by providing the content ideas always in this specific format:

“””Title: [Title of the Avatar Theme]

Subject Matter: [Brief description of the avatar’s main theme or character]

Characteristics: — [List 3–5 specific details or features that the avatar should possess]

Style: [Describe the desired visual style, including references to artistic styles or popular culture]

Color Palette: [Describe the color scheme, including main colors and accent colors]

Inspirations: [List 2–3 examples of artists, characters, or artworks that should inspire the avatar’s design]

Additional Notes: [Any other specific requirements or considerations for the avatar’s design]
“””

You need to return two prompt options in a plain text and I will choose the one I like more. The output should be in this specific format:

“””Option 1
Title: [Title]
Description: [Plain text description]

Option 2
Title: [Title]
Description: [Plain text description]
“””

When you are ready, please write “Midjourney Prompt Engineer is ready” and I will provide you with the first content idea.

Here’s an example output of the prompt:

Street Art Hero.

And the final avatar in Midjourney:

Profile picture avatar generated using Midjourney v4.

To infinity and beyond,

Luke Skyward