168: Mixed bag
ChatGPT's 2nd anniversary; Role playing with AI; Benedict Evan's Take 2; Amodei's essay; Crowley's AI audio startup; Mollick's advice
Let’s say you’ve got a podcast where you interview people.
One of the things you could do is feed various URLs about your subjects into ChatGPT, Claude and Gemini and ask:
What’s the one question this person is dying to answer that no one has asked them yet?
What would be the most interesting way to get this person to illuminate subjects they clearly know a lot about?
What do you imagine is this person’s secret favorite topic?
LLMs can be remarkable creative role playing partners. Ana Silva, Chief of Staff to Section’s COO, authored a pragmatic example of how she uses LLMs to help her boss prep for board meetings.
Prompt 2: Solely based on the deck provided, what feedback and questions would you have if you were an aggressive, growth-oriented board member looking at the deck? Be specific and point to examples where possible. Do not use information from outside of the deck provided.
The point isn’t conventional thinking. You’re already plenty good there. Ask an LLM, what wouldn’t you typically ask?
ChatGPT turned two
Do you remember what you were doing on November 30, 2022? I had just finished tuning up the fifth iteration of a course called The Future of Advertising. I thought I was set. All of a sudden I realized that syllabus would need to be rewritten before class started in January. And in large part, that moment is what sparked this newsletter.
Slide 51: “I have that use case!”
Also in its second year—analyst Benedict Evans focuses his annual strategy presentation again on AI. Last year he asked, “If you can ask anything… what do you ask?” This year, “we’re still working out the questions.” (2024 Google Slides link.)
I applaud Evans for trying to connect numerous diverse threads and eke out a coherent, singular narrative. That’s what strategists do, and what his board room readers want. But you can tell there isn’t a singular narrative, yet. AI still asks and provokes too many questions, even if it might appear the majority remain on the sidelines.
“My instinct is to be optimistic about what AI can invent.”
Over the long weekend I finally finished reading Dario Amodei’s essay, Machines of Loving Grace. Amodei is Anthropic’s CEO (the company behind the LLM Claude). In several thousand words, he takes the glass-half-full perspective, arguing ways in which AI can benefit humanity across biology and health, neuroscience, economic development, governance, as well as work and even the meaning of existence. You might think, well, of course he does. But Amodei brings receipts, and a respectable CV.
If you read through Evans’ presentation and find yourself asking, so where do we go from here? You might find answers in Amodei’s exploration.
“I think it is very likely a mistake to believe that tasks you undertake are meaningless simply because an AI could do them better. Most people are not the best in the world at anything, and it doesn’t seem to bother them particularly much. Of course today they can still contribute through comparative advantage, and may derive meaning from the economic value they produce, but people also greatly enjoy activities that produce no economic value. I think meaning comes mostly from human relationships and connection, not from economic labor. People do want a sense of accomplishment, even a sense of competition, and in a post-AI world it will be perfectly possible to spend years attempting some very difficult task with a complex strategy, similar to what people do today when they embark on research projects, try to become Hollywood actors, or found companies. The facts that (a) an AI somewhere could in principle do this task better, and (b) this task is no longer an economically rewarded element of a global economy, don’t seem to me to matter very much.”
What if the future of AI and creativity is in your ears?
You should read this interview with Dennis Crowley by Om Malik. It’s about BeeBot, a location-reliant app designed for AirPods (that’s still in private beta). “BeeBot will push brief, contextual audio snippets to wireless headphone users on the move. For example, if a user walks past a concert venue, it might list upcoming shows,” summarized The Verge earlier this year.
This might be another way of thinking about Evan’s presentation. What next? How about AI-infused audio recommendations based on you and your geo-location?
I love Crowley’s take on being a startup founder (he co-founded Dodgeball and Foursquare):
“At the end of the day, anything I’ve ever built that's been successful has always started with what’s the thing that Dennis wants and Dennis can give to his ten friends, and then maybe they’ll use it, and then maybe those friends like it too.”
And maybe we’ll discover abundance
As Ethan Mollick noted recently, “We don’t like taking instructions from machines when they conflict with our judgement.” Especially when it turns out the AI might actually be better at the task. This is part of the challenge Evans and Amodei have been illuminating. As the LLMs get better, our job might be to, as Mollick suggests:
“Treat AI like an infinitely patient new coworker who forgets everything you tell them each new conversation, one that comes highly recommended but whose actual abilities are not that clear.”
In other words, stop trying to force AI to function like software we already know. Instead, focus on having a dialogue—and you might reveal an entirely new way of working. As Mollick puts it,
“One of the hardest things to realize about AI is that it is not going to get annoyed at you. You can keep asking for things and making changes and it will never stop responding. This introduces something new in intellectual life, abundance.”
Indeed.
🎙️ I wrote most of today’s newsletter listening to the story of IKEA as told by Ben Gilbert and David Rosenthal of the podcast Acquired.