Can AI Replace Our Graphic Designer?

Introduction

The host introduces the Dolly Challenge, an AI model that can generate images from text prompts. The challenge is to compare the results of Dolly and a graphic designer in creating composite images.

The Dolly Challenge

• Introducing the Dolly Challenge, an AI model that generates images from text prompts.

• Comparing the results of Dolly and a graphic designer in creating composite images.

Understanding Dolly

The host explains what Dolly is and how it works.

What is Dolly?

• An AI model that generates realistic images from text prompts.

• Can create an image out of any given text prompt.

How Does It Work?

• Give it text, and it will make a realistic image out of that text.

• Uses machine learning algorithms to generate images based on textual input.

The Challenge

The host explains the challenge between Tim, the graphic designer, and Dolly in creating composite images.

The Prompt

• Giving Tim and Dolly the same prompt to create composite images.

• Tim has all his resources while giving only 10 seconds for dolly to complete its task.

Tim's Thoughts

• Tim feels intimidated by the challenge as he has only 10 seconds compared to infinite time for dolly.

• Acknowledges that some of dolly's creations are impressive when they hit their mark dead-on.

Starting the Challenge

Tim starts working on his first prompt while explaining his tools and setup for image composites.

Tools Used by Tim

• Wacom Intuos Pro

• Sketchbook

• Image gallery collected from Unsplash

• Ambitious approach towards handling composites for brand work such as F150, Tesla Plaid, and Mac Studio.

Starting the Prompt

• Tim collects a few images for the first prompt.

• He is trying to figure out which ones will be most beneficial for merging two images in the prompt.

• He marks some of them with a green dot to indicate his preference.

Creating Composite Images

Tim explains how he creates composite images and starts working on his first prompt.

Creating Composite Images

• Part of Tim's job is doing image composites along with logo design and brand work.

• He has to be ambitious with how he handles these composites, which is going to be a little bit more out of his comfort zone.

Working on the First Prompt

• Tim selects an image that he had stuck in his head from the beginning.

• He duplicates it, turns it into a smart object for later use.

• Tries different compositions while keeping the same dimensions as Dolly's output.

Working with Images

The speaker discusses working with images for a project and how they can be edited and composed to fit the desired outcome.

Editing Images as Smart Objects

• The speaker found an image to work with for the first prompt and saved it as a smart object.

• This allows them to edit the layer whenever they want, such as changing the composition or cropping.

Selecting Images for Composition

• The speaker pulled several images for the second photo prompt.

• They need to find images that work best with the base image and tweak them accordingly.

• Different parts of the deer require different perspectives, lighting, and focus.

• The speaker is trying to use different birds and feathers but may have difficulty due to differences in coloration, patterns, and size.

Challenges in Image Composition

The speaker discusses some of the challenges they face when composing images for their project.

Matching Coloration on Deer's Belly

• The belly of the deer has different coloring than other parts of its body.

• The speaker needs to figure out how to get feathers on the chest and belly that match this coloring so that it looks like a deer rather than just a bird inside a deer shape.

Selling Image Composition

• It's important to pay attention to small details in order to sell an image composition.

• Wrapping everything in feathers will not look convincing; instead, it's necessary to push those little details that make it look more realistic.

Progress Update

A progress update on editing process after three hours into editing.

Structure Coming Together

• Three hours into editing, structure is starting to come together.

• Lighting is getting roughed out but still needs more work.

• What's currently on screen is further along than what was asked for in the prompt.

Working with New Prompt

The speaker discusses working on a new prompt and finding suitable images to work with.

New Prompt

• The speaker is starting a new prompt that they haven't seen before.

• They found some great photos to work with and think they can create an interesting composition.

Forcing Perspective

• The initial composition is ready, but the second image needs to be forced into perspective to fit the environment of the first image.

Cutting Out Robot Image

• The speaker needs to cut out a robot image and place it on top of a wall of computers.

• They love this photo because it looks...

Conclusion

The speaker discusses their progress in editing images for their project. They discuss challenges faced when composing images, such as matching coloration on different parts of an animal's body. Additionally, they provide updates on their progress throughout the editing process and discuss how they are working with new prompts.

Dialing in Depth of Field

In this section, the speaker is discussing how they are trying to match the depth of field with the background to make the subject feel more integrated. They also mention that they are excited about their next project.

Matching Depth of Field

• The speaker is trying to dial in the depth of field to match what the background is doing.

• This will help make the subject feel more integrated into the scene.

• The end result will be revealed to the team later.

Excitement for Next Project

• The speaker is excited about their next project, which will be a little more goofy.

• They have a better idea of what they want to do and how they want to tackle it.

• They have four hours left to complete it and are looking forward to an enjoyable edit.

Prompt Number Three

In this section, the speaker discusses prompt number three, which is supposed to look like an old Italian Renaissance painting. They talk about finding images that embody a regal and stoic look for their subject.

Renaissance Painting Aesthetic

• Prompt number three is supposed to look like an old Italian Renaissance painting.

• The aesthetic involves a regal and stoic look on the figure's face.

• The speaker has found some images that work well for this style.

Base Image Selection

• The selected image will serve as a base for prompt number three.

• Since the subject needs to be holding something, this image works perfectly.

• There are options for where the head can be looking - either at the camera or at the object.

Reaction and Judging Time

In this section, there is discussion around how challenging prompt two was and how everyone is feeling. The team judges the work of two individuals, and there is some disagreement about which one followed the prompt better.

Reaction to Prompt Two

• The speaker's face is red from the stress of prompt two.

• They are anxious to see if they are worthy of being on the team.

• There is discussion around replacing someone with a robot.

Judging Work

• The team judges the work of two individuals.

• One prompt involves a deer covered in feathers, while the other involves a robot woman guarding a wall of computers.

• There is some disagreement about which one followed the prompt better.

Tim's Painting Challenge

In this section, the participants discuss and evaluate each other's paintings in response to a challenge.

Evaluating the Paintings

• The left painting is the clear winner because it has more detail, such as the clothes rippling and the flower in the background. It also features a goat taking a picture with an iPad, which is more precise than the right painting.

• Both participants agree that Tim made the left painting.

• In the case of "Deer Covered in Feathers," Tim painted on the left while Dolly painted on the right. For "Robot Woman Starting a Wall of Computers," Tim painted on the right while Dolly painted on the left.

• The left painting was made by Tim for "Goat Taking a Picture with an iPad."

Final Thoughts

• The participants express their admiration for each other's work and plan to hang some of them up as art pieces.

• They discuss how they could have improved their paintings if they had more time or resources.