Capitalogix

To Rebirth & Spring Cleaning

Next Sunday is Easter, but yesterday was the first night of Passover - an 8-day long Jewish holiday that recounts the story of Exodus. 

The overlap can be seen in DaVinci's Last Supper, depicting a Passover Seder and Jesus's last meal before his crucifixion. 

 

 

Part of the Passover Seder tradition involves discussing how to share the story in ways that connect with different types of people, recognizing that everyone understands and relates to things differently.

To do this, we examine the Passover story through the lens of four archetypal children — the Wise Child, the Wicked Child, the Simple Child, and the Child Who Does Not Know How to Ask.

The four children reflect different learning styles — intellectual (Wise), skeptical (Wicked), curious (Simple), and passive (Silent) — and highlight how we must adapt communication to the diverse personalities and developmental stages of our audience.

This seems even more relevant today, as we struggle to come to a consensus on what to believe and how to communicate with people who think differently. 

 

On a lighter note, one of the memorable phrases from Exodus is when Moses says, "Let my people go!"  For generations, people assumed he was talking to the Pharoah about his people's freedom.  But after a week of eating clogging food like matzoh, matzoh balls, and even fried matzoh … for many Jews, "Let my people go" takes on a different meaning.

After Passover, and as we enter a new season, it's a great time for a mental and physical 'Spring Cleaning,' and delve into your experiences to cultivate more of what you desire and less of what you don't.

Here is to Spring, Re-Birth, and Spring Cleaning.

Hope you had a great weekend.

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! 

Here Are Some Links For Your Weekly Reading - October 19th, 2024

Here are some of the posts that caught my eye recently. Hope you find something interesting.

Lighter Links:

• Pope Francis and the Vatican Just Created an “AI Bible” Reshaping Faith in the Digital Age. (TheBrighterSide)
• People Think They Already Know Everything They Need to Make Decisions. (ArsTechnica)
• 8 Expert Tips for Getting Started With NotebookLM. (Google)
• Chinese Humanoid Robot is the 'Fastest in the World' Thanks to One Surprising Addition. (UniladTech)
• Copy of US Constitution Found in Filing Cabinet Sells for $9M. So, Somebody Does Still Value the Constitution. (BBC)

 

Trading Links:

• OpenAI's Advantage Over Rival AI Companies Vanishes, Report Finds. (Fortune)
• Nvidia’s Shocking Truth vs. The Dot-Com Bubble. (JoshBelanger)
• The Next Wave of AI Won’t Be Driven by LLMs. Here’s What Investors Should Focus on Instead. (Fortune)
• Amazon Announces Nuclear Power Deals as Tech Giants Scramble for More Clean Energy. (GeekWire)
• ForecastTrader – Futures-Based Prediction Market for the Presidential Election and Other Things. (ForecastTrader)

 

The Evolution of Michael Jackson

This lighthearted post has something to do with artificial ...  but nothing to do with artificial intelligence.

While doing my weekly reading and web browsing (which is how I pick those links you probably think an algorithm selects), I happened upon a post about Michael Jackson on Twitter (now called X), and I enjoyed it (or at least was drawn to click and watch it). 

I grew up a huge Michael Jackson fan. As a kid, I watched the Jackson 5 Saturday morning cartoon show. His albums were the soundtrack to my college years. Later, my first wife and I saw him in concert several times. We shared that love with our youngest son, Zach. 

It's funny to look back on, but Zach used to dance to Michael Jackson's songs on stage at his Elementary School talent shows or at random restaurants. There was no choreography ... but lots of movement. I still smile when I think about it. 

You might smile (or shake your head) while watching this short video chronicling the evolution of  Michael Jackson's face changes from birth to death. 

 

via MikeBeast

It's a staggering difference. I won't pretend to know what led him to make the changes, but they're substantial. 

That being said, his music is both timely and timeless - which is very rare. He managed to make music in each era that fit in with the times but still felt very Michael Jackson. 

 

via MikeBeast

 

Gone too soon!

Here Are Some Links For Your Weekly Reading - February 18th, 2024

Here are some of the posts that caught my eye recently. Hope you find something interesting.

Lighter Links:

• OpenAI's Product Page for Sora – Their AI Model That Creates Video From Text (with Examples). (OpenAI)
• You Are Not Late to Artificial Intelligence. (TheAlgorithmicBridge)
• ‘Mental Time Travel' Can Help You Make Better Decisions. (LifeHacker)
• ‘The Alternative' is Just the Book Economists Should Read — And Won't. (Washington Post)
• Musk Says Putin Can't Lose in Ukraine, Opposes Further Funding – “Prolonging the War Does Not Help Ukraine.” (Bloomberg)

 

Trading Links:

• OpenAI Staff Granted Permission to Liquidate Shares at Astounding $86 Billion Valuation - Insider Report. (Cryptopolitan)
• What Sam Altman's Chimerical Trillions Say About AI Hype. (Axios)
• The US-China Chip War is Escalating. (Apricitas)
• In Memory of Byron Wien, Here Is a 2024 "Ten Surprises" Tribute. (BlueFoxAdvisors)
• How Big Is the Commercial Real Estate Mess in the U.S.? Expect Losses in the Trillions of Dollars Over Time. (TheInstitutionalRiskAnalyst)

 

Some New AI Tools I've Been Hearing Good Things About

At the core of Capitalogix's existence is a commitment to systemization and automation. 

Consequently, I play with a lot of tools.  I think of this as research, discovery, and skill-building.  There is a place for that in my day or week.  However, few of those tools make it into my real work routine.

Here is a list of some of the tools that I recommend.  I'm not saying I use them regularly ... but in some cases, I should (it reminds me of some exercise equipment I have). 

Since the late 90's, I've been collecting tools to make my business more efficient and my life easier. 

It's a little embarrassing, but my most popular YouTube video is an explainer video on Dragon NaturallySpeaking from 13 years ago.  It was (and still is) dictation software, but from a time before your phone gave you that capability. 

Today, I have more tools than I know what to do with, but here are a few that keep coming up in conversations. 

Daily.AI - AI Newsletters

We've launched a new newsletter that is AI-curated by Daily.ai.  It matches the tone and style of my newsletters, gets good engagement, and (honestly) looks much better than what we put out.

Our handwritten newsletters still do better on some metrics - but it's a nice addition and a promising technology. 

For transparency, our handwritten newsletters get around 45% opens, and 10% click-throughs and our new AI newsletter is averaging around 38% opens but around 30% click-throughs.  Our normal newsletter isn't focused on links - which is why the click-through isn't as good. 

That said, Daily.AI is a great tool that creates compelling two-way communication with your audience way easier and cheaper (time, money, and effort) than something similar done manually.

This is an example of how people won't get replaced by AI ... people will get replaced by people who use AI better. 

Opus.pro - AI Video Repurposing 

Opus.pro takes your long-form video content and cuts it into short-form content that you can post as teasers to other channels like YouTube Shorts, TikTok, or Instagram Reels. 

For me, I tend to do more long-form content naturally.  I go on a podcast, speak at a conference, or to a mastermind group.  I end up with a 30-minute+ video that I don't have the time or interest in using.  This tool allows me to find the best parts quickly - and still allows editing to make it perfect. 

For someone interested in really pursuing video, it's not worth it.  If you're interested in being present online, and not worried about making it your career, this is a great tool to streamline and systematize your process. 

Type.AI 

Type.AI is an interesting AI-first document editing tool.  At its core, Type is a faster, better, and easier way to write.

A lot of people are using chat GPT for some editing.  Type is a good example of a next-generation tool that incorporates ChatGPT and other LLMs under the surface.  It is aware of what you are doing and lets you know what it can do. 

Type jumpstarts the creative process and banishes the blank page.  Underneath its gorgeous UI are powerful features for generating ideas, querying your document, experimenting with different models, and easily formatting your work.

The point is that with a tool like this, you don't have to be good at prompt engineering.  The tool does that for you – so you can focus on the writing.

GetVoila.AI 

Voilà is an all-in-one AI assistant in a browser extension.  That means it goes with you everywhere you go on the web and supercharges your browser by making it use ChatGPT to do what it does best without you having to do more than check what is available.

 

Voilà simplifies the process of working with the content of websites and URLs, making it easy to convert them into various types of content, summarize them, or extract key information.  For example, if I am watching a YouTube video, I can right-click and choose "Summarize" - it creates a short video summary from the transcript.  Or, while writing this sentence, a simple right-click lets me choose: Improve, Fix grammar, Make longer, Make shorter, Summarize, Simplify, Rephrase, or Translate.

 

This is more useful than I thought.  I use it well ... then forget about it.  When I use it again, I often find that it got better.  I think you will find that with a lot of the tools these days.

 Don't worry about how well you use tools like this.  It is enough that you get better at using tools like these to accomplish what you really want.

Explore a little.  Then, let me know what you found worth sharing.

The Beatles New Song ... And Where AI Art Fails

Have you listened to the new Beatles song?  It took almost 50 years and new technology to create.

How did this happen?  An AI system, made by Peter Jackson, uncoupled the vocals from the piano on a poor-quality tape demo from the 70s.  The result – a song that would have never seen the light of day was able to bring John Lennon back from the dead to release new music for a new generation. 

You can listen to it here.

 

via The Beatles

Was it a touching tribute and closure to an extraordinary legacy?  Does it qualify as AI "art"? 

We are seeing a surge in creativity due to the rise of generative AI. 

People are doing amazing things with AI ...  and it's making entrepreneurship accessible to a new group of people.

AI is exciting, but it is also scary.  I would argue that it is a net positive.  However, there are also clear drawbacks (and potential risks).  For example, there are the obvious ones like deepfakes, art being stolen and fed into models without consent, etc.  But, there's one that many aren't talking about...

It's a lack of nuance or understanding of art. 

Here is an example of using generative AI to improve a famous art piece.

 

@Soncharm

In my opinion, the creator completely missed the point when they tried to improve Nighthawks by Edward Hopper. 

The truth is that I don't know their intent or thought process.

However, Nighthawks is famous for a reason.  It shows a patron, presumably at a late-night diner, with a desolate urban streetscape outside.  To the right audience, it embodies the isolation of a 24-hour modern society and big cities, and the hidden changes of the 20th century.  It is a poignant composition and one of the most famous American art pieces. 

Lost in the power of Generative AI, this X (Twitter) user transformed the piece by running it through an art processor multiple times. 

First, he had AI parse the image and write a description of it.  Then, he had it regenerate the image from the description.  The setting became light.  He thought he could make it better, so he moved people outside.  He parses a description again and creates a new image again.  He did this several times. 

The result is what you see — a beautifully created composition lacking any depth. 

The AI did its job; the human did not. 

A better prompt or a more artful process would have had a better result. 

But is it art?

Once created, art is in the eye or mind of the perceiver.  So, should we care who or what creates it?

Here are some other questions worth pondering.  Is AI at its best when it's amplifying human intelligence - rather than replacing it?  Or ... is the goal simply to amplify intelligence?

The Universe often gives you increasingly painful chances to learn a lesson. 

What do you think we're supposed to take from this?

Swedish Inspired Algorithms

Some Professors put together IKEA-inspired instructional booklets for their algorithms and data-structures lectures. The idea was to make easy-to-understand explanations by removing words, and only using images. Ideally, this would allow them to be understood regardless of their native language or culture. 

They've put together a plethora of assembly instructions, including for Quick Sort, BOGO Sort, Public Key Crypto, Binary Search, Merge Sort, Balance Tree, Graph Scan, and One Stroke Draw algorithms. 

via ideainstruction

This is a pretty cool idea, or at least I thought so. My youngest son said, "I don't particularly understand IKEA directions or algorithms .... so this is basically the worst of both worlds for me."  Finally, we agree about something!

Hopefully, you find it helpful. If not, there's always Wikipedia. 

Another AI Art Gimmick

There's a new trend of AI-generated profile pictures - using tools like avatar.ai.

Below are some sample results my son, Zach, got after downloading an iPhone app called Lensa.  The company that makes it also offers a product called Prisma that is worth looking at as well.  He paid $4 within Lensa to generate these AI photos (for context, avatar.ai charges more than 5x as much).  He uploaded 17 photos from different angles (and with different expressions), and the app promised 50 pieces from 4 styles.

What do you think of the results?  I thought some of them were surprisingly good. 

 

Some were not so good ... (or at least outside my zone of artistic preferences).

After seeing his, I'll admit I got curious and made my own. 

 

To be honest, these apps are still volatile ... meaning, sometimes they are good enough to seriously impress – but they also can underwhelm from time to time.  Check out DALL-E for more examples of the dynamic range of possibilities and results.

Regardless, I am impressed with AI art's progress and momentum.  Recently, generative AI has gotten a lot of hate (along with the hype), nonetheless, people are talking about it (meanwhile AI art is still not as polarizing as Kanye). 

The reality is, most of these tools still are an elegant use of brute force.  Nonetheless, I am bullish.  Hardware and software are getting better ... and when fed more data, generative art depictions get better as well. 

The uses continue to get more elegant and complex as time passes ... but we're still coding the elegance. 

For more on this topic: 

• Dall-E and AI-Generated Art 
• AI editing underwater photography
• AI evaluates how normal your face is
• On deepfakes and fake news
• Bruce Willis and AI movie characters

The Swish Machine

Rube Goldberg machines impress me more often than not. Here's a 70-step outdoor machine that covers a lot of ground ... all to put a basketball in a hoop. It took a month to create, and another month to get working. 

 

via Creezy

I love stuff like this because it reminds me of life. Looking backward, you see how all the pieces go together. As you're going through it, it feels random and sometimes like you're moving backward or that your effort isn't directly contributing to your goal.

Nonetheless, it all comes together in the end.

Onwards!