Evolving Daily Habits in the Wake of Exponential Innovation

Written without the assistance of AI. Image generated by DALL-E.

A little over a year ago, I began wrestling with a question about how businesses could create defensible value in the wake of exponential innovation. This question was prompted by the exciting and sometimes overwhelming advancements made in Generative AI throughout 2023. If you didn’t read it, you can here.

In the post, I shared five views:

1. The cost of general intelligence is going to zero. Businesses should focus on delivering value deeper into customer workflows.?
2. Businesses that can disseminate internal knowledge rapidly will differentiate themselves.?
3. Proprietary data advantages and workflow complexity are king.
4. Redefining areas of focus throughout the hiring process will become critical.
5. Great communicators and great leaders will separate themselves.

You can find supporting detail and my thoughts behind each view in the post itself. Surprisingly, I think these have aged pretty well. Looking back, I’m more proud of point 5 than I was when I wrote it. In what Bob Ross might refer to as a “happy accident”, I separated “great communicators” and “great leaders” in a way that distinguishes them from each other. With AI’s ability to bridge individual gaps, I think that point is more meaningful than I intended it to be. We could easily see a rise in leadership amongst groups that were previously disadvantaged or perceived as not being great leaders because they weren’t great communicators.

Switching modes, at the time I’m writing this, the end of 2024 isn’t that far away. Time has passed and a lot of innovation has occurred throughout the year. My personal and professional views have evolved. More than ever, writing my thoughts down and sharing them feels important to me. Whereas in the earlier post, I shared thoughts related only to business, this post straddles personal and professional. Somehow, I’ve found myself on some pretty deep text message exchanges with friends and colleagues asking for some of my thoughts, whether that’s related to parenting or writing code. It turns out, a lot of people are thinking about this. Whether this ends up being more for myself, or for others, nonetheless, I wanted to jot some thoughts down.

And, before I get too far, let me emphasize that these are my personal views. I’m wrong, a lot. So, take this all with a grain of salt. I’m thinking out loud. I welcome your critique, though.

In 2024, my daily life materially changed because of AI.

While AI has been a large part of my life for a while, I can point to 2024 as the year my daily habits changed because of it. I’ll unpack how, below.

I think differently.

When it comes to my “every day” state of mind, I’m thinking deeply about the future of AI and what it means for my family. I didn’t feel the need to do this before. Overall, I’m unsettled. I swing from being extremely excited, to not excited at all––projecting nostalgia into the future. I imagine explaining to my kids about how we used to create software and businesses by writing code with our hands, typing into a text editor all night and refreshing our browser to check the results. We’d buy programming books from Barnes & Noble and literally RTFM. Linters and autocompletion were game changing innovation. We’d wake up in the middle of the night to boot servers that had crashed, reading logs, patching memory leaks and talking to tech support at the data center for hard reboots. There will be a whole new wave of “uphill, both ways” stories to tell kids soon. It is funny, but, so much of my personal fulfillment came from those moments. The challenges and problems we solved back then continue to look smaller and smaller as much of it can be solved automatically today. It’s bitter sweet. And, it’s weird. At the same time, some of my friends are choosing to ignore the progress around them. Some take swipes at people trying to innovate, and startups that are swinging and stumbling as they try to learn where the next frontier lies. They proudly circulate examples of LLMs making mistakes, claiming that Generative AI is nowhere near being ready to augment them, let alone replace them. I feel, in many ways, like I’m living in two worlds. I worry that they’re walking examples of the truest form of cognitive dissonance. Simultaneously, my head races at all of the potential and the opportunity that we have. What can’t we build? We can and should be aiming higher. The opportunities for technical, ambitious, and creative people have never been greater. I feel a new sense of urgency, responsibility, and overall excitement.?

I learn differently.

The change in the way I think has driven a change in the way that I learn. I’m using LLMs to unpack math-heavy portions of white papers on Arxiv , to help me truly understand what’s going on. I’ve switched the majority of my data “intake” from reading to listening. This year, I figured out that I can learn much faster by listening, thanks to Google’s NotebookLM project . My spend on Audible has surpassed my spend on streaming services. I’m on Spotify every day. I’ve forked a python repository that converts PDFs to audiobooks so that I can multi-task. It seems obvious in hindsight, but, when your primary mode of learning is reading, you limit when and where you can learn. Learning, for me, has become more mobile and continuous in 2024.

I code differently.

These days, work doesn’t have me writing much code, but I code in the little free time that I have. This time last year, about 30% of my code was generated by Github Copilot in Visual Studio Code. I’ve now migrated to Cursor , a fork of VS Code, while using Claude 3.5 Sonnet by Anthropic instead of Github Copilot. Now, more than 80% of the code that I write is generated for me. Very rarely do I do anything but speak into Cursor and review Claude’s work. I moved away from using Apple’s Siri for text-to-speech and I bought a subscription to BetterDictation , which wraps OpenAI’s model, Whisper. Last year, I used TTS about 0-5% of the time, I use it about 20-30% of the time now, instead of typing. It bothered me for months that, for the most part, I no longer write code by hand. Some of the most intellectually rewarding experiences of my life have revolved around working on challenging problems, and writing code to solve them. I’ve always loved engaging and getting lost in “deep work”. The kind of work where you start with your coffee in the morning, blink, and realize it’s time for dinner. For months, I wondered if that was an outdated experience that I shouldn’t expect to have again, knowing that Claude could just write the code for me. I took comfort in learning that mathematicians felt the same way in the 70s during “The Calculator Wars”. Back then, mathematicians believed the mass use and distribution of pocket calculators risked undermining the integrity of mathematics, and they felt like it was losing its magic. It stung for them, just like this did for me.

I’ve since been able to find deep work again, not through typical hands-on programming or software engineering, though. I can’t justify doing that manually anymore, and neither can my friends or any software engineer that claims to be a good steward of its company’s or client’s resources. However, there’s plenty of room and opportunity for deep thinking at the system and architectural levels. Software engineers need to elevate themselves to system engineers, now. These days, I spend more of my time thinking deeply about system design and educating the LLM on the standards I’d like it to follow by building out documentation and prompting it to measure its work against my expectations and designs. I provide guidance regarding which libraries to use, and how to use them. I view LLMs more like point-solutions better suited for tasks, whereas I provide the direction, end-to-end, and systems-level thinking. LLMs aren’t there, yet. Not consistently, at least. I’ve wondered why they aren’t there, and how long this will be true. I’d be surprised if this were still true in 3-5 years. I think the primary reason why we’re so much stronger at systems-level thinking than LLMs is mostly due to a cost and hardware issue that’ll eventually be solved. Most of us are proactive and continuous thinkers. Right now, AIs are reactive thinkers. Except for some exceptions, people think fairly continuously. Our wheels spin, even at rest. We lay in bed thinking about the day and what’s happening tomorrow. Our dreams are fueled by the things we’ve seen throughout the day, or thoughts in our subconscious. We process all the things we’ve seen, heard and reflected on, and we spontaneously connect dots and have “Aha!” moments. AIs don’t have subconscious layers or “Aha!” moments. They simply react to the direction we provide and use statistics to complete a task in the moment. I call our ability to connect dots “systems-level” thinking. We identify two or more related concepts and bring them together into a functioning system or thought, sometimes intentionally, sometimes not. There’s a massive pile of value waiting to be unlocked by people who combine their ability to think in systems with the rapid level of execution that AI can provide. This should be an area of focus.

I invest differently.

Most of my startup investments are tech-oriented. I put a pause on investments for most of ’24 as I refreshed my theses and revisited what excites me. Things that I used to be interested in, but that no longer interest me, include categories like Developer Tools and Productivity Apps. That’s not to say they won’t do well, I just don’t have an interest in them anymore. I believe for companies to have an outsized impact on the future, their vertical depth and overall system complexity need to be very high. Developer Tools and Productivity Apps, and the like, from my perspective, often fall short in at least one of these two areas. I’m also not staying up at night getting excited or thinking about how badly software engineers need better tooling, or that productivity behind the screen is the most exciting and pressing issue of our time. I see AI eating up a lot of these opportunities.?

Solutions that run on or acquire proprietary data that’s hard to get, products that span multiple complex software or hardware systems, or technology that requires a fusion of software and hardware, are particularly interesting to me now. I believe there’s a new frontier of opportunity developing at the intersection of digital and physical innovation. Companies are being built at this intersection now that will be recognized as great over the next 5-10 years. Having businesses like SpaceX front-and-center is encouraging people to think bigger, at a time when we’ve never been better equipped to do bigger things.

A while ago, I shared a thought that I had about developing products with longstanding value at work. I presented a mental model that analogized humans and caloric burn to AI and the consumption of compute. It centered around “Optimal foraging theory”. In ecology, this is where animals, and humans, continuously work to maximize the energy we gain from food relative to the energy we spend acquiring it. In other words, most of us could walk or run to the grocery store to get a steak, but instead, we buy cars and gasoline to drive there. We spend money to preserve energy because it’s more efficient and not doing so comes at an opportunity cost. All of this has changed the lens by which I view investments. We should ask, “Is this business developing a complex enough solution that returns enough value, that, even if AI could replicate it, it would choose to buy vs. build?”. Extreme examples of categories that fit this mold are things like nuclear, advanced manufacturing, and supply chains. Interestingly enough, to my earlier point, these are areas that rely on the fusion of digital and physical innovation. In a future where very capable AIs may become your primary buyers and users, complexity and value need to be there, otherwise your users will build vs. buy.

I parent differently.?

A friend of mine frequently reminds me that the human brain is not wired for the exponential. This makes grappling with exponential innovation particularly hard. Complicating this, we’ve been “wired” for survival, for so long. This means that while incredibly positive change may be happening at exponential rates all around us, if any one particular thing arises that challenges our worldview or our personal situation, it becomes all-consuming and we ignore or miss the positive change. Politics, anyone?

A major focus of mine as a Dad is making sure my kids can always find the “good”, while cultivating their ambition and interest in creating and building as much as possible. We need our kids to be filled with ambition and optimism that motivates them to acquire the capabilities to drive change. That type of foundation is timeless. Mister Rogers said, “When I was a boy and I would see scary things in the news, my mother would say to me, ‘Look for the helpers. You will always find people who are helping”. This is a great example of a parent teaching their child to find the good. Like Mister Rogers’ Mom, my Mom (and Dad) instilled values in me that have given me advantages that I try not to take for granted. When I was on summer break one year, I was lazy and hanging out on the couch all day. My Mom walked over to the couch, looked me in the eyes, and said, “Every second you sit here doing nothing, you are robbing people of your gifts”. Think about that from a kid’s perspective. Imagine someone you love more than anything saying that to you, with the deepest level of conviction, and what that does to your your sense of purpose, your self-confidence, and your self-worth over time. It’s an incredible advantage that I was gifted, and I have the responsibility to make the most of it, and to pass it along. Ambition and creativity have been, and will be key to finding and attaining success in an AI-first world.

Now, the above approach is timeless and not new, so it’s a relatively safe take. However, there are some riskier changes my wife and I are making, as well. The kind of moves that, when you look back, you’ll either be glad you did it that way, or, you’ll wish you’d taken a different approach. Below are two of these areas.?

First, there’s our approach to teaching. Academically speaking, a lot of what we teach in schools and the way that we teach makes very little sense now. It will make even less sense in the near-future. Because our education system is relied upon to do so much more than educate our kids (e.g. we lean on it to feed, shelter and protect so many of our children during the day), it’s extremely difficult to get our system in the US to “keep pace” with innovation and the real world. Especially now. We rely on our schools to provide Maslow’s full hierarchy. This helps and hurts. There are things we can do as parents to help this. I write stories with the kids using ChatGPT to teach and reinforce life lessons. They create designs and inventions using DALL-E. They’re learning how to interact with AI, talking to ChatGPT and Alexa, feeling out the nuance of human-to-AI interactions vs. human-to-human interactions. After all, natural language is the new universal programming language; This is a critical skill to develop. When it comes to homework and practice, LLMs have infinite levels of patience. When my son successfully counts by 3s to and from 99, but suddenly cannot answer the question, “What is 21 minus 3?”, rather than me blowing a gasket, we can spin up a lesson with the help of an LLM to get a more patient teacher with soothing tone to go back and forth as we work it out. Far fewer tears can be shed over math homework. What’s interesting is, in that moment, your kids see more than math. They see you recognizing that you’re both stuck. They’re struggling to learn math and you're struggling to teach it. They see you articulating a problem verbally and asking for help in a way that generates results, to help each of you. Good communication and problem solving are materially more valuable than arithmetic. As Thomas Edison, who admitted to not being very good at math theory (but was clearly great at problem solving) said, “Mathematicians are a dime a dozen, and I can always hire them. They can’t hire me.”

Lastly, there’s work, and purpose. If done right, the two eventually intersect. It doesn’t usually start that way, however. You grind at something you’re dispassionate about in order to earn the opportunity to work on something you are passionate about. Work starts with the purpose of filling your stomach, but as you advance, its should evolve to filling your heart. In that early stage, ambition, your work ethic, and a vision for something better gets you through it. In the latter stage, your beliefs about your own purpose accelerate you into the “self actualization” phase of your life. Parents have been teaching work ethic for a long time now, and I have nothing nuanced to add there. It seems straight forward and as though the traditional methods still work. We make the kids do chores and bring them along with us when we do our own, etc. Purpose, is interesting, though. For people of faith, like our family, I think this is much more easily solved than those with different views. I can only speak for us. In the Christian faith, they teach that people are made in the image of God. That’s not to imply that people are Gods or can be, of course, in fact Christianity regularly reminds us how messed up we all are, and I think we all see that front-and-center each day. But rather, by teaching that we’re designed in God’s image, it implies that to our core, we’re made to be creators. Something aches in us when we sit idle and don’t feel as though we’re contributing meaningful value to the world. We feel less, “alive”. We have an internal “tug” towards a sense of purpose, and we believe that’s where it comes from. We can and are made to create life, do good works, and to love. All of us have unique gifts, and purposes, and it’s made extremely clear that we’re expected to make the most of the gifts we’re given––and, there’s no time to waste. To help our kids truly believe they’re as gifted as they are, to get them to pursue those gifts, and to believe in and pursue their purpose– we’ve made the switch to a new school system that more deeply emphasizes those very things. In decades, where work may be an optional thing for many, those without a clear understanding of their purpose can and will spiral more easily than those with a purpose. People may “opt out” of ambitious and tiring work pursuing big visions, as their basic necessities become “givens”, and VR and technology continue to advance to levels that enable people to escape reality with illusions of fulfillment and success. Yikes. You can see the pendulum swinging. Let’s get out of here.?

So, what’s next?

With so much changing, so quickly, I’d argue that one of the most important things we can do is make time to think. The demands of our personal and professional lives can make that pretty hard. I believe most generations perceived themselves as living through significant eras of change, or inflection points in humanity. Some of them actually did, whereas others projected false alarms. On one hand, I wonder if I think about these things too much. On the other hand, my hands-on experience automating things that many people seem not to be aware of yet, or opt out of, has me left with little doubt that we are actually experiencing a significant era of change. I say it a lot, “Technology has never been so far ahead of the market”. This is a pretty weird thing to experience in actuality. I’m simultaneously overwhelmed with excitement for the future and deeply convicted that there are areas that we need to begin rethinking–– so that they become opportunities, and not issues.