Capitalogix

How AIs (Like ChatGPT) Learn

I first wrote about this video a few years ago, when artificial intelligence first captured the media's attention. Back then, we were impressed that algorithms could help you pick out your Christmas gifts on Amazon, suggest new music for you on Spotify, and do their best to capture your attention on websites. 

AI is seemingly everywhere now. And what is surprising to me isn't its prevalence or impact ... it is the speed and breadth of its adoption. Last night, at an early holiday dinner, we talked extensively about how easy to adopt and how accessible tools like ChatGPT are (and almost none of us were Nerds). 

With all that being said, I think it's wise to have a basic understanding of the things that most impact our lives. Even if you're not a big fan of AI, understanding its growing powers will benefit you. 

The video below is a bit simple in its explanations, but it describes some fundamental concepts worth understanding.

 

CGP Grey via Youtube

The video is engaging and easy to understand. It focuses on genetic algorithms (which is one type of machine learning) and ignores some of the other more complicated techniques and approaches. For example

In my experience, it is more useful for an executive to understand what tools like this can do rather than how they work. Likewise, it is better to understand when to use tools like this rather than knowing precisely how to use them. But, a cursory understanding like this video still adds value. 

As machine learning gets more complicated and evolves, it gets harder for humans to assess their output or process accurately. Here is something to consider:

How do you know that the answer it gives is the answer you seek?

Just because an algorithm responds quickly and confidently doesn't mean it's right.

While bots can deliver impressive results, their decision-making processes can be opaque. This presents both a risk and an opportunity. Artificial intelligence seems cool, but artificial stupidity is scary ... and making mistakes at light speed rarely results in good outcomes.

It's human nature to feel safer when we understand something. It's human nature to envision machines making human-like decisions, just faster. But we are quickly going beyond that ... way beyond that!

In the past, algorithms were static while data changed. But now we're in a different world - one where the algorithms themselves evolve and dramatically adapt to handle different types of data. While this might sound like a subtle distinction, it represents a fundamental shift in how AI systems learn and operate.

One of the challenges of understanding exponential technologies is that their progress isn't linear. This makes it difficult for humans to accurately gauge how rapidly these tools will advance in capability. As algorithms grow more complex, they will increasingly operate in ways that may not be fully understood by their developers (and even less so by their users). As a result, we'll likely find ourselves using AI to solve problems and accomplish tasks that aren't even on our radar today. 

It might sound strange, but it doesn't matter why a bot makes a decision or what inputs it uses to make the decision. What matters is whether it accomplishes its goal and how its performance and level of decision-making rank in relation to its prior performance and other options (and, perhaps, whether the bot is biased).

Part of what makes artificial intelligence exciting is that it can do a lot of things well that humans are really bad at. And, even when you're using an AI in your domain expertise, it can be a great first step to save you time and effort. 

It's a brave new world, and not only is Big Brother watching, but algorithms are, too.

Live long and prosper!

Why Leave The House When You Can Shop Online ...

The holidays are almost here! Time's running out to get your gifts in time. 

Luckily, it's never been easier to get stuff. On top of online shopping, Amazon has stewarded the switch to almost instantaneous shipping. 

I remember when malls were filled to the brim, and Black Friday led to stampedes and people dying. Now, it's brick-and-mortar stores that are dying.

Not only are the in-store deals not as good, but the best deals are online. Here's a chart showing the surge of online shopping.

 

via visualcapitalist

Online shopping now accounts for about 15% of consumer spending, up from 4.1% in 2010. 

In addition, Q4 always sees a massive spike, with online spending accounting for 17.1% of consumer spending in 2023. 

This year, Cyber Monday attracted over 64 million U.S. shoppers, three times the number who shopped in stores. 

You can find change everywhere, but one place to look hasn't changed. Amazon is all about infrastructure and disruption. 

• Amazon announced Rufus, a new generative AI-powered conversational shopping experience.
• Amazon announced a way to get your prescriptions online faster and cheaper than a traditional pharmacy.
• Now, you can even buy cars on Amazon.

It's interesting to consider how drastically this changes supply chain management. People now expect free and fast shipping and the ability to return items if they don't work out (instead of trying them in-store). Meanwhile, brands have high fixed costs due to their existing retail locations, yet they need to keep most of their stock in warehouses instead of their box stores to deal with the surplus of online orders. 

We're already seeing the failure of malls and the bust of commercial real estate. If this trend continues of work-from-home and online shopping, we could see a $250 billion decrease in commercial property value by 2026.  

What do you think it would take to stop the bleeding and encourage both consumers and employees to leave their homes?

Ready Fire Aim!

Michael Masterson wrote a book called Ready Fire Aim: Zero to $100 Million In No Time Flat. It is a practical guide for entrepreneurs and business leaders, focusing on the different stages of business growth and the key challenges and priorities at each stage.

A core message of the book is to start taking action quickly (instead of getting bogged down in over-planning) and use rapid iteration and real-world feedback to refine strategies.

The concept is presented in the context of growing a business – yet the lessons apply broadly. 

Swift action should be your focus ... not over-planning or perfect timing.

Too many companies get stuck in a cycle of brainstorming, getting internal feedback, making changes, and failing to release the product.

Even for released products, too many fail because they took too long to launch or ignored market feedback. 

Masterson stresses that the value of live performance is that it helps you course-correct. He also cautions that there is no such thing as perfect timing. The best timing is almost always ‘Now!’

Now, let’s extrapolate. 

Let’s say you’re pondering a tricky work problem. You know you need to figure it out before the end of the week ... but your brain keeps going in circles. 

You don’t believe you can take decisive action and course-correct because you feel you have to get it right. 

So, what can you do? Write it out. Write out the potential paths, ramifications, and worst-case scenarios as holistically as possible. 

You’ll find that simply by writing it out instead of just ‘thinking,’ you end up more creative with more insight. 

Writing aids in organizing and clarifying your thoughts. 

AI As Your ‘Action’ Partner 

AI can make this process easier, faster, and more manageable. Tools like ChatGPT can help you explore complex topics, run scenarios for you, and provide external feedback. 

It’s not the same as real-world feedback, but it can shorten the ‘Aim’ part of the process. 

As you externalize your thoughts in writing them to whatever AI you choose, you also get the same benefit you did writing your thoughts out in our previous example. 

The business environment is changing faster than ever. Technology is advancing faster, adoption is getting easier, and the average Joe is becoming bullish on AI. Because of that, Masterson’s book is even more relevant now than it was in 2007. Here are a few of the other key takeaways from Ready, Fire, Aim.

• Adapt Your Role: As a company scales, leaders must shift from “doing” to guiding and from tactics to strategy.
• Be Sales-Focused: Keep revenue generation as a core priority, especially in the early stages.
• Build Scalable Systems: Invest in operations and leadership at the right time to support sustainable growth.
• Innovate Continuously: Avoid complacency by fostering an organizational culture that balances efficiency with creativity.

AI and exponential technologies are going to compress cycle types. What used to be long-term planning will just be planning. You have to act fast, not only to capitalize on these trends but also to avoid being wiped out by them. 

Are you keeping up?

Finding The Path Of Least Resistance ...

There’s a concept in design and transportation called Desire Paths. 

A Desire Path is the path users take instead of the path intended by the builder. 

Here’s a great example. 

Reddit via itstartswithani

If you are interested, there is an active online community forum that shares examples of Desire Paths. It may give you some ideas and knowing laughs.

I am a creature of habit, and even though much of what I think, feel, or do seems to be happening based on real-time choices or decisions, much of that is just a well-worn rut of unconscious behavior.

As a subtle reminder to my son, who just got married, expect many of your existing Desire Paths to change (even if you don’t want them to).

The lesson ... It’s often easier to account for or take advantage of human nature (or nature) than to fight against it. 

Here is a short video on how this relates to your business and tech adoption. I call it Functional Mapping. Check it out. 

 

The video provides additional depth and detail beyond what’s covered in this post. I encourage you to watch it for a more complete perspective.

Understanding the natural path for both technology and people makes it easier to understand and anticipate the capabilities, constraints, and milestones that define your path forward.   That means you actually have to understand the different types of users and what they expect to do. Here’s a diagram that explains how we build AI-enabled applications.

Each stage is really about the opportunity to scale desired capabilities and automation.

It isn’t really about building the technology; instead, it is about supporting the desire.

You don’t have to get it right. You just have to create momentum in the right direction.   Meaning ... if you can anticipate what is coming, you don’t have to build it. Instead, you should figure out where you want to build or create something that will move things in the right direction to help make that happen or benefit from it when it happens.

You’ve probably heard me talk about how Capabilities become Prototypes. Then, Prototypes become Products. And, ultimately, Products become Platforms.

This model is fractal. That means it works on many levels of magnification or iteration.

What first looks like a product is later seen as a prototype for something bigger.

SpaceX’s goal to get to Mars feels like their North Star right now ... but once it’s achieved, it becomes the foundation for new goals.

This Framework helps you validate capabilities before sinking resources into them. 

It helps you anticipate which potential outcomes you want to accelerate. Rather than simply figuring out the easiest next step … you have to figure out which path is the best next step to your desired outcome.

The world is changing fast! Hope you’re riding the wave instead of getting caught in the riptide!

Onwards.

How Hard Is It To Run A Business In America?

Running a business is hard. Of course, most people know that – but now VisualCapitalist has put together a chart to help visualize that sentiment. 

 

via visualcapitalist

20% of companies tracked failed in the first year ... and only one-third last a decade.

The numbers are worse than that in tech, mining, and oil & gas. Interestingly, those industries receive some of the largest government subsidies. 

But, if you are reading this, chances are that you are a survivor. So, take pride in knowing the odds you've overcome ... and be thankful. 

It is Thanksgiving this week.

Onwards! 

The World Is In $100 Trillion Dollars of Debt ...

The world is swimming in debt ... well, to be more specific ... the world's governments are swimming in debt — $100 Trillion of it. 

via Barrons

To put that in perspective, here's an illustration to give you a sense of the enormity of that number.

via Barrons

The U.S. accounts for just over 34% of that number. Meanwhile, I remember writing about the Republican National Convention marking the moment our national debt crossed the $16 Trillion level in 2012. 

To put the current number in context, if our national debt were divided among individuals, we'd each owe more than $100K ... and if the ten wealthiest people donated their entire fortunes, we'd only have covered about 5%. 

The concept of "Debt" can be confusing to a layman. Most people understand what it means when they take on debt with a local bank, but it can be harder to understand the role debt plays in global economics.

Compounding the confusion, the implications of debt change on a macro level. 

Many worry that our "excessive" government debt levels impact economic stability, the strength of our currency, and unemployment. The national debt can only be reduced through five mechanisms: increased taxation, reduced spending, debt restructuring, monetization of the debt, or default. 

The idea behind our current global debt structure is that if two nations are mutually obligated and dependent on each other, they are less likely to go to war. And that has held relatively true so far. Of course, it's not a perfect system (and it could break down), but it's working better than previous systems (such as the balance of power).

In some ways, it's fake money, so our debts don't seem insurmountable or fatal. Our economy is so reliable that we're allowed to continue borrowing. Debt is an integral part of the economic machine - it can be argued that we wouldn't have money without debt. 

Ray Dalio created a simple (but not simplistic) and relatively easy-to-follow 30-minute animated video that answers the question, "How does the economy really work?"  Click to watch.

via Ray Dalio

The global economy has grown enormously during the last 50 years as developing nations prosper. The average global GDP per capita has gone from ~$1000 to over $10,000 in my lifetime.

So, it makes sense that the amount of debt is also increasing with the size of the money supply required to conduct all the transactions in the global economy.

But, even though you may not need to be immediately worried about that number, I still think it's worth trying to put it in context. 

Humans are notoriously bad at large numbers. It's hard to wrap our minds around something of that scale. We're wired to think locally and linearly, not exponentially (it's one of the reasons I love AI so much). Here are a couple of ways to help you understand a trillion dollars. 

via AskOpinion

First, let's look at spending over time. If you were to spend a dollar every second for an entire day, you would spend $86,400 per day. If you have a million dollars, you can do that for approximately twelve days. With a billion dollars, you can do that for over 31 years. With a trillion dollars, you can do that for 31,000+ years. That means it would take over 300 thousand years to spend the global public debt at that rate. 

I'm sure many of you make over six figures a year. But, it would still take you 10 million years - if you spent none of it - to make $1 trillion, let alone $100 trillion. 

Let's try explaining it through time. Fifty thousand seconds is just under 14 hours. A million seconds was 11 days ago. A billion seconds ago from today? 1992. One trillion seconds is slightly over 31,688 years. That would have been around 29,679 B.C., which is roughly 24,000 years before the earliest civilizations began to take shape. Pretty crazy. 

Here's a video from the 1970s that helps you understand scale through the power of tens and an exploration of our universe. 

Eames Office via BetterExplained

Hopefully, that was disturbing and helpful!

Futurism and The Epidemic of Impossible Statistics

I can’t pretend this is a new phenomenon, but I also can’t pretend it’s not becoming a pet peeve of mine. 

If you’ve been following me for any amount of time, you’ll know I love the future, and I love random statistics. If I’m not talking about AI or entrepreneurship, it’s generally because I’m sharing some interesting chart or statistic. 

At the intersection of my two loves comes a pretty severe issue ...

Bullshit statistics. 

Futurists can’t help themselves. If you repeat something enough times, it begins to feel true. This is a key part of the reality distortion field that surrounds charismatic leaders. Their “functional fiction” becomes useful - not because it’s grounded in fact, but because it enables us to envision what’s possible and work to make it real.

In their defense, nobody minds if you talk about the future broadly. However, a problem arises when directional belief masquerades as fact or science. For example, if someone has thought about something many times, there is a tendency to confidently discuss or project exponential growth with specific timelines and metrics (rather than broadly discussing what will eventually come).

This tendency can make intelligent people seem delusional (or at least out of touch).

Elon Musk is a great example. While he has undeniably been a significant force for innovation and progress in the world, here are a few of the outlandish claims he’s made recently.

• “I think by 2040, probably there are more humanoid robots than there are people.”
• “I feel comfortable saying that AI is getting 10 times better each year.”
• Or, his 70% chance of moving to Mars, and his expectation of colonization starting in 2029.

Some of those may be true, and all of them might turn out to be right ... but they are still wild-ass guesses.

Elon is by no means the only one doing this. 

I routinely make up statistics to help me simplify or understand things better. The key is to acknowledge these “shortcuts” are still essentially educated guesses. Here’s an example. When I imagine how advanced AI will become by the end of my lifetime, I have to consider my current age (and expected lifespan) and how rapidly AI is improving. If I assign the number “100” to how good AI will be at the end of my life, what value would I assign to it now? Turns out, I’d give it a value of three. Of course, there’s always the possibility I could get hit by a bus tomorrow. I’m not a scientist. I haven’t done detailed research about chips or when we move to quantum computing. Realistically, I don’t have to. The precise numbers aren’t what matters here. I don’t take that statistic literally. It’s directional, and it gives a sense of the rate of change and the velocity of invention. In that sense, even though it isn’t factual, it’s useful.

I’d say any serious scientist knows that you can’t reliably predict the future with that level of precision – but it doesn’t take a scientist to know that. 

First, the statistic or shortcut has to pass a simple “sniff” test. Then, you have to account for likely bottlenecks or constraints. Too many of these crazy estimates assume almost infinitely scaling results with no setbacks or limitations in materials or energy. 

Don’t underestimate the value of a good rule-of-thumb or mental model. Moore’s Law is a great example of that. It stemmed from an observation and prediction about the semiconductor industry made by Intel co-founder Gordon Moore in 1965. A grossly simplified version is that computing power doubles every two years. That has held true for more than 50 years.

I have two Gaping Void illustrations that express fundamental truths about this: “First, Bring Order to Chaos” and “Wisdom Comes from Finer Distinctions.”

Here’s the reality. The future is exciting ... and it’s coming fast. In many ways, it will likely be bigger and cooler than you could have imagined. In other ways, it will radically underperform your expectations. 

I can say that not because I know any more than you, but because I’m focused on what doesn’t change. We’ve had many periods of innovation ... each bigger than the last. It’s likely there will be aspects of the next 20 years no one can predict. But, we know what innovation looks like. 

We’ve been here before. 

As a reminder, if it sounds too good to be true ... there’s a good chance it is. Yet, to pretend there’s not a chance outlandish claims will come true would be to make too precise a claim again. 

In many ways, predicting how your business or product will change is much easier than how the world will change.

The best way to predict the future is to create it - and the most effective way to create it is to focus on the elements within your control.

While it’s important to play an exponential game ... you can start “locally”. 

Food for thought! 

How do U.S. States Compare to Entire Countries?

With Trump's election win, the economy will likely be center stage. 

Crypto boomed post-election, and the S&P and the Nasdaq reached new highs as the Fed cut rates. 

Amidst all the political turmoil, it's easy to overlook how well-positioned we are as a country – compared to the rest of the world. 

VisualCapitalist created a chart comparing countries' GDPs to those of U.S. states. 

 

via visualcapitalist 

At the top of the list, California would be the world's fifth-largest country (replacing India) with its $3.9 Trillion GDP. 

In second place on the list, my home state, Texas, had a $2.6 Trillion GDP. For a little more context:

• If Texas were a country, its GDP would fall just behind France ($3.1 Trillion) and above Brazil and Italy ($2.3 Trillion).
• The Texas GDP is also higher than Canada's $2.24 Trillion GDP.

For a bit of contrast (and extra perspective), Vermont has the lowest GDP in America, comparable to Nepal, which is the world's 101st-largest economy. Yet, Nepal has 29 times more people than Vermont. 

Not too shabby for a place where dairy, forestry, and maple syrup drive the economy.

The NFL's Halftime Hack

When you think of food meant to help athletes, you might think back to when parents would give orange slices or Gatorade to kids during soccer games. Now, fast forward your thinking to the best options available to fuel today's performance athletes. What do you imagine they would use before or during a game? 

I quickly went to protein bars or energy gels (like GU) ... But, I was wrong.

According to research from The Athletic, The NFL's favorite halftime energy booster is now the uncrustable.

What is that? An "uncrustable" is a crustless sandwich consisting of a filling between two layers of crimp-sealed bread. They are convenient, portable, and they tend to come in homey flavors like PB&J. While you can theoretically make one at home, Uncrustables, in this context, are made by Smuckers. 

The Athletic via X 

Based on this math, it's reasonable to assume that NFL teams go through ~4000 Uncrustables a week (on top of however many PB&Js or other sandwiches they consume). Regardless, it is a lot. 

Why the sudden surge in popularity? It's a combination of factors:

• Nostalgia and Comfort: The uncrustable taps into a sense of childhood nostalgia, offering a comforting and familiar taste.
• Nutritional Balance: It provides a balanced blend of carbohydrates for quick energy, fats for sustained power, and protein for recovery.
• Convenience: Pre-packaged and mess-free, uncrustables are perfect for a half-time snack when you have very little time to eat. 

If you think about it, it makes sense. PB&Js are universal. Regardless of geographic region, socioeconomic status, etc. – it's a staple. 

They're easy to digest, convenient, and comforting. Hitting all the marks a professional athlete might be looking for. 

Even in a high-tech world, people still crave simplicity and effectiveness.

I've never had one ... but now I kind of want to try. On the other hand, I'm pretty sure they used to call this a Pop-Tart. But I guess that just makes me seem old.

Getting Educated Before You Vote!

Early voting has started, and if all goes to plan, we’ll know who will be the President on Tuesday, November 5th. Realistically, it’s more likely that you’ll know who you think the President is ... but the country may still be officially undecided.

It has gotten more challenging to exercise truly educated decision-making through the fog of fake news and the growing echo chambers of noise surrounding the candidates (and their policies).

The reality is that most of our news sources are tinted by the lens of their ideology - which is okay. That means they’re appealing to their targeted audience. 

That being said, it may also be worth looking at the same news story from a comparative news source from the other spectrum - or even better, a primary source when it comes to science/economics. I found two tools that help me do this kind of comparison easily – Ground News and Particle.

Ground News is a news aggregation platform that makes it easy to compare news sources, read between the lines of media bias, and break free from algorithms. Key features and benefits include:

• Source comparison: Allows users to see how different news outlets cover the same story, grouped by political leaning (left, center, right)
• Bias distribution: Shows the breakdown of political orientation for news sources covering each story
• AI-generated summaries: Provides summaries of coverage from different political perspectives
• Blindspot feature: Helps users identify stories not widely covered by their preferred sources
• Ownership information: Shows the ownership/financial interests behind news outlets.

Particle is a news reader app that delivers AI-powered news summaries to help you understand more, faster. It was created by an ex-Twitter alumni. The Particle app lets you browse a personalized list of stories to get you up-to-speed at a glance ... or go infinitely deeper to understand various perspectives, broader context, and how a story has unfolded over time. They’re still in private beta. You can sign up for the waitlist here.

Here’s a chart that shows where news sources rank. You can click the image for an interactive version with more details. And, if you’re curious about their methods, click here. 

via Ad Fontes Media

On top of reading various news sources, there are a couple of other tools I like: 

• FiveThirtyEight - Uses graphics to tell visual stories on valid statistics. Named after the number of electoral votes. Also has a visual presidential forecast. 
• RealClearPolitics - Aggregates news from various sources, as well as writing their own opinion pieces. Also has a good visual on the current state of the presidential, senate, and house races. 
• Brittanica ProCon - Tracks the stated positions of politicians on various issues. It can be sorted by candidate and by issue. It also has a quiz you can take to assess which candidate you actually resonate with. 
• Politifact & FactCheck - Both sites fact-check presidential statements, party statements, and more. We know politicians often lie by omission or focus only on the specific stats that are relevant to their point. Fact-checking helps you gain a more holistic picture. 

You are probably sick and tired of all the noise surrounding the election. However, our democratic system thrives on the intricate system of checks and balances powered by information sharing and continuous debate. This process compels the government to remain responsive and accountable to the people. 

Remember, choosing not to vote is still a choice ... and getting out to vote is the only way to make sure that your voice is heard - whether or not you believe in the health of America's democracy.