06 June 2023

A.I. Design Wars: Why Adobe Is Going To Crush MidJourney (#126)

2Frameworks

10Insights

Frameworks in this episode

Strategy6 steps

Platform-versus-Point-Solution AI Test

Judge AI products by workflow breadth, retention power, distribution, and defensibility.

Read →

Marketing5 steps

Rapid Creative Feedback Loop

Ship low-cost creative tests quickly and let real audience feedback guide iteration.

Read →

Insights & moments

The myth-busts, hot takes, explainers, and tools worth keeping.

Myth Buster· 1

Myth Buster06:30

AI Commoditizes Average Skill, Not Exceptional Expertise

The hosts reject the idea that universal access to AI makes professional expertise irrelevant. AI may bring many users into an acceptable range, but exceptional practitioners will still outperform because they understand the underlying discipline and can use the tools more effectively.

AI gives many users similar baseline capabilities.
Average production skills may become commoditized.
Deep disciplinary expertise remains differentiating.
Experts can extract more value from the same AI tools.

“there's always gonna be truly great people who use AI tools better than anyone else because they just truly understand that discipline of work.”
Kieran Flanagan · 06:30

#expertise#skill commoditization#creative work

Hot Take· 4

Hot Take01:00

AI Images and AI Video Help Different Skill Levels

Kieran argues that image tools can raise weak or average creators toward competent output, while the best AI video results still tend to come from people who already understand video. The technologies may therefore democratize quality at different rates.

Image generation already produces useful results for ordinary users.
AI video tools can look artificial when used without production expertise.
Experienced video creators use AI to amplify existing craft.
Different media have different AI learning curves.

“I think there is somewhat opposite ends of the spectrum where images like, really for that not really great to okay bucket, and really moving…”
Kieran Flanagan · 03:00

“the people who are already good great at video, and they're producing like incredible stuff.”
Kieran Flanagan · 03:30

#ai images#ai video#creative skills

Hot Take16:30

AI Training Copyright Rules Could Become a Geopolitical Advantage

The hosts discuss Japan's permissive position on using copyrighted material for model training and argue that such rules could influence where AI companies operate. Divergent national approaches may reshape the competitive advantages of incumbents and startups.

Training-data law affects model development.
Permissive jurisdictions may attract AI businesses.
Copyright policy could become geopolitical competition.
Regulatory changes may erase a startup's data-access advantage.

“Part of this whole image war that's happening is the copyright law and the copyright infrastructure of the world.”
Kipp Bodner · 17:00

“It's going to become a geopolitical issue, because there are gonna be some countries that are gonna come down on different sides of the copyright…”
Kipp Bodner · 17:00

#copyright#japan#ai regulation#geopolitics

Hot Take18:00

Discord Is Midjourney's Mass-Market Bottleneck

The hosts praise Midjourney's technology and community but argue that its Discord-based experience creates unnecessary friction. To reach mainstream users, it needs a dedicated interface outside a service most of the world does not use.

Discord helped Midjourney build an early community.
The interface is unfamiliar and frustrating for many users.
Early-adopter enthusiasm does not guarantee mainstream fit.
A standalone interface could expand adoption.

“Mid Journey has to get off Discord. They're gonna be irrelevant if they don't get off Discord.”
Kieran Flanagan · 18:30

“The majority of the world is not in Discord for better and for worse, you know?”
Kipp Bodner · 19:00

#midjourney#discord#user experience#adoption

Hot Take19:30

The Image Wars May Be Won by a Universal AI Assistant

Adobe, Midjourney, Canva, and Figma may not be the ultimate winners if users prefer a single assistant that handles search, images, video, and other tasks. If models commoditize, control of the unified interface could matter more than ownership of a specialized model.

Users may prefer one interface for many creative tasks.
Bing already combines search and image creation.
Model commoditization could weaken specialist differentiation.
Universal assistants could disrupt established creative platforms.

“the true winner is just this conglomerate of an AI assistant that can do all things.”
Kieran Flanagan · 19:30

“do you just have a Google-esque conglomerate in the future where you can just do all things through a single interface?”
Kieran Flanagan · 20:00

#ai assistants#creative platforms#model commoditization

Explainer· 2

Explainer03:30

AI Compresses Years of Creative Learning Into Days

Kipp distinguishes the internet's democratization of information access from AI's democratization of understanding and execution. People may soon express ideas in music, design, and other media without first mastering years of technical instruction.

The internet expanded access to information.
AI helps people understand and apply information.
Creative ideas can become artifacts without years of technical training.
Image-based design is an early beneficiary.

“the internet democratized access to information. Artificial intelligence is democratizing understanding of information, right?”
Kipp Bodner · 03:30

“make that learning curve days not years”
Kipp Bodner · 04:30

#democratization#creative learning#ai

Explainer09:00

AI Can Estimate Creative Success Where Historical Data Exists

Kieran explains that creative performance becomes more predictable when AI can access a relevant history of outcomes. YouTube tools can analyze concepts, titles, and descriptions, but novel creative work without comparable data will continue to produce substantial hits and misses.

Historical performance data can inform creative predictions.
YouTube provides a rich source of comparable outcomes.
AI can estimate the probability that an idea will perform.
Novel work remains difficult to forecast without relevant evidence.

“anywhere where you have creativity paired with data and data sources, AI is gonna start to be able to tell you what is the percentage…”
Kieran Flanagan · 09:30

“anywhere where it's hard to pull like actual data to tell me historically what works or does not, that's where there's a lot of more…”
Kieran Flanagan · 09:30

#predictive analytics#youtube#creative data#ai

Takeaway· 3

Takeaway04:30

Generative AI Gives Designers Better Starting Material

Generative tools improve not only final outputs but also the raw material available to expert designers. Instead of adapting generic stock photography, creatives can begin with relevant generated imagery and build stronger advertisements or websites from it.

Traditional stock assets are often generic or weak.
Generated images can match a specific creative need.
Experts can use stronger inputs to produce stronger finished work.
Everyday visual communication may become more relevant.

“we're also just starting now with much, much higher and stronger and more relevant baseline design content.”
Kipp Bodner · 05:00

“being able to create better baseline, like aspects of a work for expert designers and everything to take advantage of is a big net positive.”
Kipp Bodner · 05:30

#graphic design#generative ai#creative assets

Takeaway08:00

Creative Speed Makes Cultural Timing a Competitive Advantage

AI can prevent a timely idea from becoming irrelevant during production. As everyone gains faster execution capabilities, marketers must become more capable of identifying and acting on cultural moments while they still matter.

Production delays can destroy the relevance of a good idea.
AI shortens the path to a finished creative asset.
Faster competitors will increase audience expectations.
Marketers must learn to operate in the moment.

“by the time those couple of days go by, it's no longer relevant”
Kipp Bodner · 08:30

“it is gonna be more important than ever to operate in the moment in the culture”
Kipp Bodner · 09:00

#cultural relevance#marketing speed#real-time marketing

Takeaway22:00

AI Makes Every Marketer Broadly Capable but Not Equally Valuable

AI lets marketers perform a minimum level of design, video, and other specialist work, making broader T-shaped capability more common. The differentiating question becomes where each person can deliver tenfold depth while using AI to cover adjacent tasks.

AI provides baseline ability across multiple disciplines.
Broad competence will become common among marketers.
Individuals still need an area of exceptional depth.
Integrated AI workflows increase shipping speed.

“we can now all do like a minimum amount of everything because AI can help us do it.”
Kieran Flanagan · 22:00

“where am I 10x, right?”
Kieran Flanagan · 22:30

#t-shaped marketers#career skills#marketing workflows