Seven Questions for AI

Future technology is an age-old mirror of humanity's present

Welcome to the latest entry in The Captain’s Log. You’re in good company with thousands of fellow entrepreneurs and innovators who have subscribed!

I’m your host, Bob, and my mission here is to share personal, behind-the-scenes stories of ups and downs from my career leading tech startups and corporate innovation.

I write to make you think, smile, and discover a shortcut to success or a trap to avoid.

Here we go…

I paid for one of those AI headshot generators. This is the only version that wasn’t embarrassing to look at. Plus, I always wanted my own action figure.

This is one of my favorite weeks of the year. It’s an annual gathering of about seventy diverse thinkers who guess the future of business for a living. Coburn Ventures, an amazing, outlying advisory group, organizes the event. These sessions typically include a fair share of investment analysts, entrepreneurs (that’s me!), inventors, artists, and scientists. And you’ll often find a priest, a monk, and a rabbi. That’s no joke.

There are no booths to visit, no clients to woo, and no keynote speeches. Instead, there are fast-moving breakout discussions where everyone is encouraged to share their unique perspectives. Nearly everyone ends the day with a new collection of notes and friendships.

This year’s topic is Artificial Intelligence. It's no surprise, of course, that a group of thinkers on the future need to pool their brains on this one. We’ve been asked to bring our current beliefs on where it’s going. But as I began to organize my thoughts, I realized that I have more questions than answers…

I don’t mean questions about what will happen in the market but rather what’s happening in other people’s heads when they think about AI. Maybe it’s the marketer in me who was trained to ask “Why?” five times to get a deep insight about my customer.

Whatever the reason, I do this all the time now. I’m fascinated by why the person in love with ChatGPT finds it so fascinating—and why the person worried about human extinction is so concerned. Why is that writer using it to challenge their posts before they publish? Why does the salesperson for this recruiting app I’m testing reference his software’s AI every 15 seconds?

So this year, I’m heading into the session with just enough opinions to add value to the discussion—but mostly, I will be asking questions. As I put my list of questions together, I thought they might help you, dear reader, cut through the fear/hype and be more comfortable with not having all the answers…

1. Why are some people so excited about AI?

I’d bet we all have people in our lives who gush with praise over the latest AI tools available. They’re building bots, designing logos, and sharing demos of all the amazing stuff that can be done.

I’m fascinated by their fascination. The other day, someone forwarded a demo of the latest OpenAI tool. In 73 seconds, it researched trend articles on fashion websites and mixed them up to create a novel shoe design. The person who shared is not a shoe designer—in fact, not a product creator of any kind. Yet they took this demo and lept to four huge conclusions about how this tool will change our lives immensely.

Every new technology brings out the dreamers and optimists, of course. But are these people really using AI as much as they say they are? Or is this similar to sex—lots of talk but a lot less action than we imagine?

Some have a profit motive to get others to act. Investors, for example, are in the business of backing founders who want to build the next next thing. So, it's no surprise that VCs like Marc Andreessen and James Currier are pumping up the troops. We, founders, are the waves of cannon fodder they count on to hit their 10x fund goals. Sam Altman has bet his career and life savings on OpenAI’s success, so he’s got to keep heaping on the hype.

Investors always get excited—they are in the future business. But what about the present generation of youngsters that will lead us into the future? We can often look at the new tools and trends they embrace to see what’s coming next. I’ve got two young 20-something daughters, so I’m slightly closer to this generation than most people. And so far, their silence on the topic of AI is deafening. Where are the Gen-Zers who are using AI tools that we old fogies don’t understand?

Then again, the overall buzz still is palpable, and change is happening. I mainly think we’re just curious creatures that love to wrap our minds around new stuff. And some of us like to be the ones who show the cool toy to our friends. It’s harmless if you don’t get too caught up in things. And as long as your mind doesn’t turn to panic.

2. Why are some people so afraid of AI?

For every person who proclaims the next great leap in civilization, there seem to be two people who fear it will end civilization forever. These “doomers” love to rain on the hype parade. And they are pulling out all the stops.

Read any anti-AI piece, and you can play a game of Fear Bingo. Squares in the game include: Comparisons to nuclear war and global warming, the number of jobs that will be lost, trigger words like “hallucination” and “black box,” concerns that more power will accrue to white male tech-bros, and news stories that take one error and blow it up to ridiculous proportions.

I think I’ve got some answers to this one:

• Our human evolutionary programming errs on fearing the new and imagining what could go wrong. Bad wins over good. Websites get more clicks from scary stuff. Scary stuff is more socially contagious.

• It’s easier to imagine bad things happening than good things that haven’t been created yet. And our collective unconscious includes decades of “tech gone wrong” stories from years of Sci-Fi movies. We readily conjure up images ranging from Skynet to the DJ3000.

Of course, every fearful story has a counter-fear argument. If we don’t push the boundaries of AI, the Chinese will. If we don’t use AI, we’ll never find and destroy a rogue comet due to hit us in 25 years.

3. Does AI help or hurt thinking?

I heard another interesting “success case” for AI the other day. A senior consultant at a big-name firm told me that she had asked a junior associate on her team to write a summary of ten articles. The associate forgot to do the task, so she asked ChatGPT for the output. Despite a few errors she saw in digging in, she was impressed by the result. It left her wondering why the +$150k/year associate was needed…

Now, let’s put aside a few things before I go on. First, consultancies need to invest in training associates over time so that they are ready to take over the big roles one day. Second, consultants bill by the hour. Both combine to make it in a company’s interest to allow some bullshit work.

I question whether turning two hours of reading into a five-minute summary helps one’s thought process.

I can’t speak to how others make decisions or develop strategic client recommendations. All I can speak to is my process, which relies on significant time invested into understanding the world and making important decisions. Working through dense research papers or speaking with seven experts for an hour each can be brutal and take significant time. But by marinating myself in the subject matter over hours, days, and weeks, my mind can make connections and develop a solid path forward.

Then again, I must live with my business decisions…but consultants (ahem) usually don’t.

4. Are we fooling ourselves?

Last year, our team launched SantAI - the first use of ChatGPT for children aged 1 to 92 to get in touch with the big guy before Christmas Eve. We’ve seen thousands of children spend an average of 23 minutes chatting with Santa—as well as more than a few adults who ask questions that definitely put them on the naughty list.

Seeing kids ask for toys, peace, and a chance to be with loved ones is touching. ChatGPT does an amazing job with this assignment. If you squint, you’ll think “he’s” real. I’d put “him” up against any mall Santa. And sometimes it’s fun to be fooled—to get in the spirit and smile.

But I wonder if we’re making leaps of logic because we want to believe AI is something more.

A few weeks ago, with friends at a bar, an AI fan in our group wanted to demonstrate a new use case. He fired up chat ChatGPT and asked it to “summarize the writing of (yours truly) Bob Gilbreath.” After a second or two of processing, it rattled off a few sentences that everyone in our group agreed were pretty accurate. Of course, it didn’t even say what I wrote about—nor why anyone would need such a use case.

Then it hit me. Our minds were looking for evidence that it was accurate. It’s the same reason horoscopes and tarot card readers seem believable. And it’s why personality tests continue to be used. I recently learned this is called The Barnum Effect:

“a common psychological phenomenon whereby individuals give high accuracy ratings to descriptions of their personality that supposedly are tailored specifically to them, yet which are in fact vague and general enough to apply to a wide range of people.”

There’s also a lot of evidence that we’re fooling ourselves in the financial markets. From newly-born startups to Microsoft/Google/Amazon, we see massive hype around AI today. Coming fresh off the Web3+NFT+Crypto run, one would think we’re all a little more wary of the hype. But maybe we’re just looking for greater fools.

5. Is AI for Chores or Creation?

I have to admit that I get giddy with excitement when I think about how AI could take over a myriad of annoying, brainless tasks that I have to do every day. I would pay anything for a Spreadsheet bot or a Shopping bot. So much work is not being done or done poorly because it is simply brain-crushingly tedious.

So why are so many people excited about getting ChatGPT to write books and rap lyrics?

Now, I can see the pull for help in creative-projects-which-involve-chores tasks, like picking an image for your substack post. I’m a paying user of Grammarly, which addresses my countless errors in writing posts like this.

I suspect people who wish to apply AI to truly creative endeavors don’t understand the creative process. It’s hard…which is kind of the freaking point. IBM Watson won Jeopardy, and Google’s AI won Go because both games allow for limited actions with known answers and repeating patterns. That’s machine learning in a nutshell.

But these machines perform poorly where they do not know the problem, they are playing in an open game world of infinite strategies and where real creativity is required. I bet the latter sounds a lot like your job. You can relax for now.

As a long-time writer, I laugh whenever I read something created by AI. It’s so obvious and so awful. At best, it is predictable—which is exactly what the software does: predict which word can go next based on the average of all other words. My one beef with Grammarly is that it increasingly wants to change my writing to take the life out of it. I suspect the company feels pressure to load new AI features to keep up with the hype.

No, bad Grammarly, I meant to use that word in an original way that will make my reader smile.

Actually, I have a second beef: The software regularly misses basic stuff that I still have to go back and check for. Are Grammarly’s product wizards moving too quickly to the next AI toy rather than fixing the small stuff that Microsoft Word could catch 20 years ago? Or are my expectations too high? (see #6, below)

[But I digress…do AIs digress? While they may dream of electric sheep, I doubt they digress. It’s not predictable.]

I have been taking guitar lessons for seven years. Once in a while, I think about writing a song and end up putting a couple of chords and words together. But songwriting doesn’t come naturally to me and is not on my priority list. But what if an AI tool could turn my few words and chords into an original tune?

I couldn't care less. It wouldn’t be a song from my soul, written through hours of pouring my heart into it.

6. Will we demand perfection from AI?

Human intelligence is far from perfect, as we know. Many hope that artificial intelligence will help us address our imperfections. We see this today in tools that can automatically improve the photos you take—including those of yourself. These are low-stakes, low-cost tools that make the world a little bit better.

But what about when the stakes are higher? Take autonomous cars, for example. This is a very high-stakes situation with real lives in the balance. We should embrace AI-driven cars because they are much safer than human-driven ones. But they, too, make mistakes.

And when those mistakes happen, people tend to react with great outrage. If your town witnessed a fatal car self-driving car accident today, the city would immediately freeze all autonomous vehicle operations until a major investigation was performed. There would be protests in the streets. Yet there will be a fatal car accident in your town this afternoon, and it might make the daily newspaper. Are we willing to look at the bigger picture of net lives saved, or expect perfection?

Special interests are already raising the stakes in AI such that even “simple” solutions become painfully debated. The optimists rush to see that ChatGPT could lead to better web searches—but the naysayers need proof that search results won’t include any trademarked phrases, trigger words, or political biases.

Startup innovations need some freedom to screw up and learn from mistakes. A of steam engines blew up before we got to versions that lasted more than a few hours. Some people lost their lives in the process. But eventually, the engine was perfected, and countless lives were improved and saved. In global governments, there is a creep toward the Precautionary Principle, which is guaranteed to reduce accidents and stunt innovation. Will we allow this technology to go through needed growing pains—or will we keep it in a padded crib? Again, it’s easier to fear and imagine the negatives.

7. Where is the Present impact?

At the end of the day, comparisons to the past or predictions about the future are meaningless—because neither exists. History is rewritten by the victors and spinners. The future is impossible to predict accurately. Humanity only exists in the present. Perhaps that’s where we should spend more time…

Look around today for real changes brought on by AI. Which companies are rising or falling? What company shut down because ChatGPT exists? Has Nike laid off its entire design team because of the 73-second shoe creator? As Cal Newport says: “Filter for tangible impact.”

Also, look for pockets of opportunity in the present—where the future is here, just not evenly distributed. As an owner of a recruiting business that works with early-stage startups, I’ve gotten an inside peek at some killer companies that have been using AI to solve real problems, and they’ve been doing it for years. These companies have developed technology to do new things like analyze health scans for disease, block intruders from software, and map aging sewer systems (talk about a job no human wants to do!).

Or you might choose to take the complete opposite path and build a business that is free of AI. As nearly every publication, podcast, and blog poster embraces an always-on AI content strategy, I have a feeling there’s less competition in all the other places where business and life happen. There will be an inevitable counter-reaction to a bot-driven world. For our businesses, it’s one of the reasons we’re doubling down on service businesses. People can do amazing things.

Finally, can we please ease back on the advice to “make testing GPTs part of our daily routine”? I had an MP3 player a few years before Apple entered the game. The battery sucked, could only hold ten songs, and the software to add and remove those ten songs was barely above DOS 1.0. Using that old device did not help me “get ahead of the competition” when the iPod and iTunes were released. If anything, I was too slow to get on board because I was “invested” in that old product.

We’ve got enough things to do in our lives—right here in the present. Most of us are not software developers; most of the tools out there look great in a demo but aren’t a pleasure to work with. If OpenAI, Apple, or XYZ Startup launches the no-brainer version of AI, we can be the first to line up.

Conclusion

You could run this post through ChatGPT and ask it what I really think about AI today. But you’re already smart enough to know that I’m on the doubtful side. Don’t get me wrong, I’m hopeful and optimistic for the long run, but I’ve lived through enough technology change by now to be a lot more realistic in the short run. Everything new goes through Hype Cycles. It takes time to evolve into something truly useful.

I’m unsure how my new Question-Asking Strategy(TM) will work out as I navigate this week’s event and the months and years ahead as a leader of my businesses. I find myself turning a question back into the mirror: Is this a wiser model of looking at the future, or am I just becoming the old fart who is wary of change? I’m excited to find out.

Try it yourself: The next time you hear someone speak with excitement or fear about AI—or about anything, actually—hold your opinion back and ask them why they think that way. That’s when things get really interesting.

---

How we might work together…

• My team and I lead Hearty, a boutique recruiting service that helps tech-forward companies hire proven talent. Our senior team of operators sources and screens, saving you time and money. When you need help, let’s chat.

• Looking for Influencer Marketing and Content Creation? The team from our previous company is back by popular demand with A2 Influence. We’re ramping up now and would love to share more.

• Feel free to schedule time together during my Open Hours for questions, feedback, networking, or any other topic!

• Find me on LinkedIn or email.

---

BONUS: Cool Content of the Week

A little something I found meaningful. You might agree…

Why VC-Backed DTC Startups and M&A Have Failed

Reader Elizabeth “Smalls” Miller turned my attention to this post from Ben Cogan about the fall of much-hyped and some-loved DTC brands:

“VC-backed DTC companies can exit in two ways: go public, or get acquired. The public brands have performed poorly. Unfortunately, the acquired businesses have if anything done worse.”

Cogan, an active investor and acquirer in the space, praises Native deodorant purchased by P&G in 2017. Native seems to be a rare winner on every measure. Likely because, “Native was profitable and cash generating at the time of acquisition. It had to be because it raised almost no venture capital.”