16 March 2023

How To Protect Your Business From GPT-4 in 2023 (#101)

1Frameworks

8Insights

Frameworks in this episode

Strategy5 steps

AI Business Defensibility Audit

Audit your data, distribution, community, and speed before AI erodes your edge.

Read →

Insights & moments

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

Hot Take· 2

Hot Take17:30

Most People Will Use Only Three to Seven AI Applications

Kipp argues that AI adoption will follow earlier media and technology markets: people encounter many options but concentrate their attention on a small number of recurring use cases. This concentration could produce a handful of disproportionately large winners.

TV, radio, social media, and mobile usage concentrated around a few choices.
Downloading many applications does not mean using them regularly.
Individuals and businesses will likely settle on three to seven AI use cases.
A small number of providers may capture most market value.

“each individual person or business, it's really gonna have like the 3 to 7 use cases. It's not going to be 50. It's not going…”
Kipp Bodnar · 18:00

“There's gonna be a couple of really large winners who have all of the weighted success.”
Kieran Flanagan · 18:30

#market-concentration#ai-apps#user-behavior#platforms

Hot Take28:00

Large Companies Have Overspecialized Their Employees

Kipp recounts encountering a job at a large company with such a microscopic scope that he was surprised it could occupy an entire role. He argues that organizations have overspecialized knowledge workers and should help people develop broader knowledge and context-switching ability.

Large organizations can fragment work into extremely narrow roles.
Excess specialization may reduce adaptability.
Broader knowledge helps employees switch contexts.
Smaller AI-enabled teams may spend more time on substantive work.

“we've overspecialize people in a lot of ways. And we need to get people back to a broader set of knowledge.”
Kipp Bodnar · 29:00

#specialization#knowledge-work#organizations#careers

Explainer· 2

Explainer10:00

Why Chat Interfaces Will Replace Many Clicks and Taps

Kipp predicts that software interaction will shift from purely graphical interfaces toward a blend of graphical, conversational, and command-based controls. Kieran supports the claim with Zapier behavior showing that OpenAI users connect and use a broader range of applications.

Users currently navigate software through repeated clicks and taps.
Conversational interfaces can translate intent directly into actions.
Chat may make complex applications easier to combine.
The shift could reshape design and user-experience work.

“A part of that usage, and I think a big part of that usage is going to move to a chat-based interface”
Kipp Bodnar · 10:00

“chat is becoming much more the kind of user interface of choice to go do a bunch of thing in all these other apps.”
Kieran Flanagan · 11:30

#chat-interface#software#user-experience#automation

Explainer25:00

Revenue per Employee May Become the Defining AI-Era Metric

The hosts expect AI-first businesses to operate with smaller teams and potentially higher revenue per employee than legacy incumbents. They propose tracking deeply AI-integrated companies against non-integrators to determine whether employee efficiency improves while profit remains stable or grows.

AI-first companies may require fewer employees.
Smaller teams can be more agile and spend less time on administration.
Revenue per employee allows comparisons across different company sizes.
Efficiency is meaningful only if revenue and profit outcomes hold up.

“My thesis is that the future of our society is way more businesses created with fewer employees.”
Kipp Bodnar · 26:30

“Revenue per employee, that is the golden metric.”
Kieran Flanagan · 26:30

#revenue-per-employee#efficiency#startups#metrics

Story· 1

Story05:00

Synthetic Models Could Replace Traditional Product Photo Shoots

Kieran describes an AI-created modeling agency capable of generating unique virtual people for online-store imagery. The example shows how generative media can collapse the cost and coordination involved in hiring models, arranging clothing, and conducting physical shoots.

AI can generate unique virtual models for individual campaigns.
Synthetic imagery removes casting and photo-shoot logistics.
Creative-service agencies face disruption beyond software markets.
Poorly rendered hands remained an obvious early limitation.

“you can replace all of these agencies overnight who book real people to do photo shoots and things like that, with AI avatars that are…”
Kieran Flanagan · 05:30

#synthetic-media#ecommerce#photography#automation

Tool· 1

Tool22:00

A Zapier AI Index Could Reveal Which Companies Are Adopting AI

The hosts propose publishing an index based on which public companies most actively use OpenAI and related tools through Zapier. They suggest that real adoption data could inform marketing analysis, venture research, partnerships, and even a hypothetical investment basket.

Zapier can observe practical adoption across connected applications.
A public leaderboard could distinguish use from corporate messaging.
The index could be divided by model or media use case.
The investment discussion is speculative rather than validated financial advice.

“the public companies that were most adopting the AI, you could just basically buy an ETF that is weighted towards all those companies.”
Kipp Bodnar · 22:30

“what you guys have is the real adoption data.”
Kipp Bodnar · 24:30

#zapier#index#adoption-data#investing

Takeaway· 2

Takeaway12:00

ChatGPT Compressed Technology Adoption From Years to Months

A comparison of major technologies shows ChatGPT reaching 100 million users in two months, while the internet took roughly four years. The hosts argue that this adoption speed means businesses should expect meaningful AI innovation in weeks rather than years.

ChatGPT reached 100 million users in about two months.
The internet took roughly four years to reach the same scale.
Digital technologies can diffuse faster than physical products.
AI strategy must account for compressed innovation cycles.

“Like the internet took about four years to get to 100 million users. ChatGPT took two months.”
Kieran Flanagan · 13:00

“the innovation in ChatGPT is not going to happen in years: it's going to happen in weeks.”
Kipp Bodnar · 13:30

#adoption#chatgpt#innovation#technology

Takeaway33:00

The Gap Between AI Beginners and Experts Is Unusually Small

Kieran argues that the current gap between novice and expert AI users is smaller than it will be later. The hosts frame rapid learning as a rare opportunity for individuals to transform their careers during a major technological transition.

AI expertise is still broadly accessible to newcomers.
The knowledge gap will widen as the field matures.
Learning speed can create career leverage during platform shifts.
The hosts compare the opportunity with earlier social, mobile, and Web2 transitions.

“the difference between an AI beginner and an AI expert has never been so small and will never be so small. It will only grow.”
Kieran Flanagan · 33:00

“it is literally just about who can learn faster than others.”
Kipp Bodnar · 33:30

#learning#careers#ai-skills#opportunity