Business

  • A Few Graphs On The State Of AI in 2023

    Every year, Stanford puts out an AI Index1 with a massive amount of data attempting to sum up the current state of AI. 

    Last year it was 190 pages … now it's 386 pages.  The report details where research is going and covers current specs, ethics, policy, and more. 

    It is super nerdy … yet, it's probably worth a skim. Here are some of the highlights that I shared last year. 

    Here are a few things that caught my eye and might help set some high-level context for you. 

    Growth Of AI

    Screen Shot 2023-04-15 at 3.38.32 PM Screen Shot 2023-04-15 at 3.13.09 PM

    via 2023 AI Index Report

    One thing that's very obvious to the world right now is that the AI space is growing rapidly. And it's happening in many different ways.

    Over the last decade, private investment in AI has increased astronomically … Now, we're seeing government investment increasing, and the frequency and complexity of discussion around AI is exploding as well. 

    A big part of this is due to the massive improvement in the quality of generative AI. 

    Technical Improvements in AI

    Screen Shot 2023-04-15 at 3.22.00 PM Screen Shot 2023-04-15 at 3.15.06 PMvia 2023 AI Index Report

    This isn't the first time I've shared charts of this nature, but it's impressive to see the depth and breadth of new AI models. 

    For example, Minerva, a large language and multimodal model released by Google in June of 2022, used roughly 9x more training compute than GPT-3. And we can't even see the improvements already happening in 2023 like with GPT-4. 

    While it's important to look at the pure technical improvements, it's also worth realizing the increased creativity and applications of AI. For example, Auto-GPT takes  GPT-4 and makes it almost autonomous. It can perform tasks with very little human intervention, it can self-prompt, and it has internet access & long-term and short-term memory management. 

    Here is an important distinction to make … We're not only getting better at creating models, but we're getting better at using them, and they are getting better at improving themselves. 

    All of that leads to one of the biggest shifts we're currently seeing in AI – which is the shift from academia to industry. This is the difference between thinking and doing,  or promise and productive output.

    Jobs In AI

    Screen Shot 2023-04-15 at 3.14.09 PMScreen Shot 2023-04-15 at 3.30.24 PMvia 2023 AI Index Report

    In 2022, there were 32 significant industry-produced machine learning models … compared to just 3 by academia. It's no surprise that private industry has more resources than nonprofits and academia, And now we're starting to see the benefits from that increased surge in cashflow moving into artificial intelligence, automation, and innovation. 

    Not only does this result in better models, but also in more jobs. The demand for AI-related skills is skyrocketing in almost every sector. On top of the demand for skills, the amount of job postings has increased significantly as well. 

    Currently, the U.S. is leading the charge, but there's lots of competition. 

    The worry is, not everyone is looking for AI-related skills to improve the world. The ethics of AI is the elephant in the room for many. 

    AI Ethics

    Screen Shot 2023-04-15 at 3.39.46 PM Screen Shot 2023-04-15 at 3.37.46 PM

    via 2023 AI Index Report

    The number of AI misuse incidents is skyrocketing. Since 2012, the number has increased 26 times. And it's more than just deepfakes, AI can be used for many nefarious purposes that aren't as visible.

    Unfortunately, when you invent the car, you also invent the potential for car crashes … when you 'invent' nuclear energy, you create the potential for nuclear bombs. 

    There are other potential negatives as well.  For example, many AI systems (like cryptocurrencies) use vast amounts of energy and produce carbon. So, the ecological impact has to be taken into account as well.

    Luckily, many of the best minds of today are focused on how to create bumpers to rein in AI and prevent and discourage bad actors. In 2016, only 1 law was passed focused on Artificial Intelligence … 37 were passed last year. This is a focus not just in America, but around the globe. 

    Conclusion

    Artificial Intelligence is inevitable. It's here, it's growing, and it's amazing.

    Despite America leading the charge in A.I., we're also among the lowest in positivity about the benefits and drawbacks of these products and services. China, Saudi Arabia, and India rank the highest. 

    If we don't continue to lead the charge, other countries will …Which means we need to address the fears and culture around A.I. in America. The benefits outweigh the costs – but we have to account for the costs and attempt to minimize potential risks as well.

    Pioneers often get arrows in their backs and blood on their shoes.  But they are also the first to reach the new world.

    Luckily, I think momentum is moving in the right direction. Watching my friends start to use AI-powered apps, has been rewarding as someone who has been in the space since the early '90s. 

    We are on the right path.

    Onwards!

    _____________________________________

    1Nestor Maslej, Loredana Fattorini, Erik Brynjolfsson, John Etchemendy, Katrina Ligett, Terah Lyons, James Manyika, Helen Ngo, Juan Carlos Niebles, Vanessa Parli, Yoav Shoham, Russell Wald, Jack Clark, and Raymond Perrault, “The AI Index 2023 Annual Report,” AI Index Steering Committee, Institute for Human-Centered AI, Stanford University, Stanford, CA, April 2023. The AI Index 2023 Annual Report by Stanford University is licensed under

  • Yield of Dreams: Laughing Your Way To Financial Freedom

    The movie Field of Dreams came out the year my first son was born. If you haven't seen it, it's a fantastic movie. 

    Whether you’ve seen it (or not), you might want to see Charlie Epstein’s a one-man play called Yield of Dreams. 

    I put together a quick video on why you should watch it:

    >> Click Here to check it out

    Many of the people who read my blog, or are subscribed to my newsletter, are either entrepreneurs or in the financial space. While Charlie Epstein moonlights as an actor/comedian, his day job is in financial services. He's incredibly sharp, very knowledgeable … and yes, a little quirky. 

    But that quirkiness is what makes him funny – so much so that you'll be captivated long enough to gain some real value. Charlie does an excellent job teaching people how to do practical things to ensure they have enough money when they retire to live a good life.

    More importantly, he helps you think about your mindsets and what you truly want, so you can live the life you've always dreamed of and deserved. And even though I didn't think I needed to learn anything new, I gained a ton of practical value – and you probably will too.

    As a bonus, half of the proceeds go toward supporting vets with PTSD.

    There aren't many people (or "offers") I'd feel comfortable plugging, but this is one of them. As well, many of the other people I would put in front of you (like Dan Sullivan, Peter Diamandis, and Mike Koenigs) love Charlie as much as I do. 

    via Yield of Dreams

    So, here's the part I copied from Charlie: In this one-man show you'll discover how to

    • Work less while making more than you ever have before
    • Make more progress towards your dreams in one year than most people do in ten
    • Step into the biggest, boldest and most confident version of yourself
    • Stop worrying about money and start living your dream life

    So, if any of that interests you I highly recommend you sign up. You only have a limited time to do so. 

    >> Just click here to learn more about Yield of Dreams

  • The Benner Cycle: Making Market Predictions

    When I first got interested in trading, I used to look at many traditional sources and old-school market wisdom.  I particularly liked the Stock Trader's Almanac. 

    While there is real wisdom in some of those sources, most might as well be horoscopes or Nostradamus-level predictions.  Throw enough darts, and one of them might hit the bullseye. 

    Traders love patterns, from the simple head-and-shoulders, to Fibonacci sequences, and the Elliot Wave Theory.

    Here's an example from Samuel Benner, an Ohio farmer, in 1875.  That year he released a book titled "Benners Prophecies: Future Ups and Down in Prices," and in it, he shared a now relatively famous chart called the Benner Cycle.  Some claim that it's been accurately predicting the ups and downs of the market for over 100 years.  Let's check it out. 

     

     

    Here's what it does get right … markets go up, and then they go down … and that cycle continues.  Consequently, if you want to make money, you should buy low and sell high … It's hard to call that a competitive advantage.

    Mostly, you're looking at vague predictions with +/- 2-year error bars on a 10-year cycle. 

    However, it was close to the dotcom bust and the 2008 crash … so even if you sold a little early, you'd have been reasonably happy with your decision to follow the cycle.

    The truth is that we use cycle analysis in our live trading models.  However, it is a lot more rigorous and scientific than the Benner Cycle.  The trick is figuring out what to focus on – and what to ignore. 

    Just as humans are good at seeing patterns where there are none … they tend to see cycles that aren't anything but coincidences. 

    This is a reminder that just because an AI chat service recommends something, doesn't make it a good recommendation.  Those models do some things well.  Making scientific or mathematically rigorous market predictions probably aren't the areas to trust ChatGPT or one of its rivals.

    Be careful out there.

  • Top U.S. Banks by Uninsured Deposits …

    It's been about a month since we discussed Silicon Valley Bank (SVB). But the impact is still lingering. I know friends whose money is still tied up, and we've continued to see increased coverage of banks' perceived failures. 

    Currently, there are $7 trillion sitting uninsured in American banks. VisualCapitalist put together a list of the 30 biggest banks by uninsured deposits. 

     

    The U.S. Banks With the Most Uninsured Deposits

    via visualcapitalist

    Many of the banks on this list are systemically important to the banking system … which means the government would be more incentivized to prevent their collapse. 

    It's important to make clear that these banks differ from SVB in several ways. To start, their userbase is much more diverse … but even more importantly, their loans and held-to-maturity securities are much lower as a percentage of total deposits. Those types of loans took up the vast majority of SVB's deposits, while they make up less than 50% of the systemically important banks on this list. But, according to VisualCapitalist, 11 banks on this list have ratios over 90%, just like SVB, which brings them to a much higher risk level. 

    Regulators stepped up in the wake of the SVB collapse, and the Fed also launched the Bank Term Funding Program (BTFP), as we discussed in the last article on this subject. But, it remains to be seen what will happen in the future. 

    Does the Fed have another option besides saving the banks and backing deposits? If not, market participants will start to rely on the Fed to come to the rescue, making even riskier decisions than they already were. 

    It feels like the Fed is stuck between a rock and a hard place, but hopefully, we will start to see some movement in the right direction.

  • Should We Temporarily Halt The Progress of Generative AI?

    Several high-profile names (including Elon Musk) have penned an open letter calling for the pause of the creation of models more powerful than GPT-4. 

    In March, OpenAI unveiled GPT-4, and people were rightfully impressed. Now, fears are even greater about the potential consequences of more powerful AI. 

    The letter raises a couple of questions. 

    Should we let machines flood our information channels with propaganda and untruth? Should we automate away all the jobs, including the fulfilling ones? Should we develop nonhuman minds that might eventually outnumber, outsmart, obsolete and replace us? Should we risk loss of control of our civilization? – Pause Giant AI Experients: An Open Letter

    The crux of their message is that we shouldn't be blindly creating smarter and more robust AI until we are confident that they can be managed and controlled to maintain a positive impact. 

    Artificial-Superintelligence-In-Control

    During the pause the letter calls for, the suggestion is for AI labs and experts to jointly develop and implement safety protocols that would be audited by an independent agency. At the same time, the letter calls for developers to work with policymakers to increase governance and regulatory authorities. 

    My personal thoughts? Trying to stop (or even pause) the development of something as important as AI is naive and impractical. From the Industrial Revolution to the Information Age, humanity has always embraced new technologies, despite initial resistance and concerns. The AI Age is no different, and attempting to stop its progress would be akin to trying to stop the tide. On top of that, AI development is a global phenomenon, with researchers, institutions, and companies from around the world making significant contributions. Attempting to halt or slow down AI development in one country would merely cede the technological advantage to other nations. In a world of intense competition and rapid innovation, falling behind in AI capabilities could have severe economic and strategic consequences.

    It is bigger than a piece of software or a set of technological capabilities. It represents a fundamental shift in what's possible.

    The playing field changed.  We are not going back. 

    The game changed.  That means what it takes to win or lose changed as well.

    Yes, AI ethics is an important endeavor and should be worked on as diligently as the creation of new AI.  But there is no pause button for exponential technologies like this.

    Change is coming.  Growth is coming. Acceleration is coming. Trying to reject it is an exercise in futility. 

    We will both rise to the occasion and fall to the level of our readiness and preparedness.  

    Actions have consequences, but so does inaction.  In part, we can't stop because bad actors certainly won't stop to give us time to combat them or catch up. 

    When there is some incredible new "thing" there will always be some people who try to avoid it … and some who try to leverage it (for good and bad purpose).

    There will always be promise and peril.

    What you focus on and what you do remains a choice. 

    Transformation Equals Innovation Plus Purposeful Action_GapingVoid

    Whether AI creates abundance or doomsday for you will be defined largely by how you perceive and act on the promise and peril you perceive. Artificial intelligence holds the potential to address some of the world's most pressing challenges, such as climate change, disease, and poverty. By leveraging AI's capabilities, we can develop innovative solutions and accelerate progress in these areas.

    It's two sides of the same coin. A 6-month hiatus won't stop what's coming. In this case, we need to pave the road as we traverse it. 

    We live in interesting times!

    What do you think?

  • Predicting The Future With Arthur C Clarke

    Last week, I shared a couple of videos that attempted to predict the future. As a result, someone sent me a video of Arthur C Clarke's predictions that I thought was worth sharing.

    Arthur C Clarke was a fantastic science fiction writer and a famous futurist. You probably know him as the author of 2001: A Space Odyssey.

    Here are his predictions from 1964, nearly 60 years ago.

     

    via BBC Archive 

    Arthur C. Clarke had a profound impact on the way we imagine the future. Known for his remarkable predictions, Clarke's ideas may have seemed farfetched at times, yet his thoughts on the future and the art of making predictions were grounded in reason.

    If a prophet from the 1960s were to describe today's technological advancements in exaggerated terms, their predictions would sound equally ridiculous. The only certainty about the future is that it will be fantastical beyond belief, a sentiment Clarke understood well.

    You can be a great futurist even if many of your predictions are off in execution, but correct in direction. For example, Clarke predicted that the advancements in communication would potentially make cities nonexistent. While cities still exist – in much the same way as in the 1960s – people can now work, live, and make a massive difference in their companies from anywhere on the planet, even from a van traveling around the country. Global communication is so easy that it's taken for granted. 

    As a science fiction author, some of what he wrote about might seem ridiculous today. For example, super-monkey servants creating trade unions.  Much of what he wrote about was what could happen (and to provide a way for people to think about the consequences of their actions and inactions).  As we discussed last week, humans often recognize big changes on the horizon … but they rarely correctly anticipate the consequences. 

    In summary, even though some of Clarke's predictions were farfetched, they were rooted in a deep understanding of human potential and the transformative power of technology. His ability to envision a fantastical future was not only a testament to his imagination, but also served as an inspiration for generations of scientists, engineers, and dreamers. By embracing the unknown and acknowledging the inherent uncertainty of the future, we can continue to push the boundaries of what is possible and strive for a world that is truly beyond belief.

    You won't always be 100% correct, but you'll be much closer than if you reject what's coming. 

  • Tech Over The Long Run

    Humans are wired to think locally and linearly … because that's what it took to survive in a pre-industrial age. However, that leaves most of us very bad at predicting technology and its impact on our future. 

    To put the future of technology in perspective, it's helpful to look at the history of technology to help understand what an amazing era we live in. 

    Our World In Data put together a great chart that shows the entire history of humanity in relation to innovation. 

    Longterm-timeline-of-technology

    Max Roser via ourworldindata

    3.4 million years ago, our ancestors supposedly started using tools. 2.4 million years later they harnessed fire. 43,000 years ago (almost a million years later) we developed the first instrument, a flute. 

    That's an insane amount of time. Compare that to this:

    In 1903, the Wright Brothers first took flight … 66 years later, we were on the moon. 

    That's less than a blink in the history of humankind, and yet we're still increasing speed. 

    Technology is a snowball rolling down a mountain, picking up steam, and now it's an avalanche being driven by AI. 

    But innovation isn't only driven by scientists. It's driven by people like you or me having a vision and making it into a reality. 

    Even though I'm the CEO of an AI company, I don't build artificial intelligence myself … but I can envision a bigger future and communicate that to people who can. I also can use tools that help me automate and innovate things that help free me to focus on more important ways to create value. 

    The point is that you can't let the perfect get in the way of the good.  AI's impact is inevitable.  You don't have to wait to see where the train's going … you should be boarding. 

    Onwards! 

  • Can We Predict The Future?!

    New technologies fascinate me … As we approach the Singularity, I guess that is becoming human nature. 

    Second Thought has put together a video that looks at various predictions from the early 1900s. It is a fun watch – Check it out. 

    via Second Thought

    It's interesting to look at what they strategically got right compared to what was tactically different. 

    In a 1966 interview, Marshall McLuhan discussed the future of information with ideas that now resonate with AI technologies. He envisioned personalized information, where people request specific knowledge and receive tailored content. This concept has become a reality through AI-powered chatbots like ChatGPT, which can provide customized information based on user inputs.

    Although McLuhan was against innovation, he recognized the need to understand emerging trends to maintain control and know when to "turn off the button." 

    While not all predictions are made equal, we seem to have a better idea of what we want than how to accomplish it. 

    The farther the horizon, the more guesswork is involved. Compared to the prior video on predictions from the mid-1900s, this video on the internet from 1995 seems downright prophetic. 

    via YouTube

    There's a lesson there. It's hard to predict the future, but that doesn't mean you can't skate to where the puck is moving. Even if the path ahead is unsure, it's relatively easy to pick your next step, and then the step in front of that. As long as you are moving in the right direction and keep taking steps without stopping, the result is inevitable. 

  • Generative AI Explained By AI

    VisualCapitalist made an infographic on Generative AI using explanations by generative AI … a bit of a novelty, but still a helpful infographic. 

    Generative-ai-infographic-MAINvia visualcapitalist

    Basically, generative AI refers to AI that generates new outputs based on the data they have been trained on. Instead of recognizing patterns and then making predictions, they're used to create images, text, audio, and more. 

    Please let me know about any tools you think are especially worthy (or that I might have missed). 

    With Google and Microsoft entering the space, I think you're about to see a lot of tool churn as they push redundant tools out of the market. Short-term, that'll cause a bit of chaos. Long-term, it will mean that we'll have a better diversity of tools as innovators are forced to be more creative.

  • South Park, ChatGPT-4, and More …

    This week, South Park released an episode written entirely by ChatGPT … about ChatGPT. 

    Here's a clip. 

    … And it's funny … Which is a testament to the power of ChatGPT.

    Then, even more news started to come out. 

    First, Google and Microsoft released announcements about integrating AI into their engines and services (like Google Sheets and Microsoft Office). 

    Then, the most important news came out. GPT-4 was released ahead of schedule. 

    In December 2022, I shared that GPT-3 was trained on 100x more parameters than any previous language model. Well, here's what GPT-4 looks like compared to GPT-3. 

    Screenshot-2023-03-10-113420And it is also much more robust. Which is mildly scary to think about … but also very exciting. 

    I was already very excited about what GPT-3 and the applications built on it were making possible, but this takes it to another level. 

    Though, I have to be the voice of caution and remind you – don't rely on it as your business. It is a tool to help you increase the speed or scope of a business opportunity. But it isn't mature enough to blindly rely on it to be right – or even factually accurate. GPT-4 was optimized to sound good, which means that it might hallucinate or mislead you, even when it has the correct answer in its dataset. But it will sound good, even when it isn't right.

    That being said, it outperforms GPT-3 in every single way. 

    Screenshot-2023-03-14-at-1-36-02-pm
    via ZDNet

    ChatGPT (powered by GPT-4) is now multi-modal, meaning it will take both picture and video inputs. 

    Exciting stuff … Or at least progress.  Onwards!