Funston Brief (Issue No. 15)
This month, our featured commentary dives deeper into the world of generative AI and offers several lessons that SMB CEOs can implement in their companies.
Generative AI has gripped the attention of executives, with earnings call mentions of the topic skyrocketing in recent months - rising more than 16x from Q4’22 to Q1’23, according to CB Insights. As AI-powered technology gains in popularity, AI content rights are gaining importance. Twitter, Reddit, and Stack Overflow announced plans to start charging LLMs to train on their data as their large datasets are increasingly valuable for AI developers like Google and OpenAI.
Meanwhile, Europe appears to be taking the initiative in the policy discussions around AI technology. OpenAI was banned in Italy due to GDPR concerns. What is the training data? What does OpenAI do with your input? Who owns the output? And how is training data paid for? To answer some of these questions, the Washington Post has analyzed which websites are in the training data. Furthermore, it's crucial to understand what 'privacy' means for a specific user's data. It is a little fragment in a massive collective corpus that doesn't reveal anything specific about the user, as privacy and copyright issues with LLMs are debated.
Companies are rushing to adapt. Alphabet, the parent company of Google, the inventor of the transformer technology (the T in GPT), and a leader in AI research, appears to be scrambling to catch up. Google Brain (internal team) and DeepMind (acquisition) were recently integrated; DeepMind CEO Demis Hassabis will lead the new combined organization. Adobe is working hard to disrupt itself before others do; it already has a product for previewing AI picture production and has shown what future videos might look like. Bloomberg, a financial, software, data, and media company, has created its GPT for financial services.
ACTIONABLE INSIGHTS.
Network effects drive your life: An illuminating deep-dive on the unseen hand of network effects that impact nearly every aspect of our lives. This article outlines how network effects impact nearly every aspect of your life. With that lens, it lays out a perspective on how to make the 7 most important decisions of your life. Hopefully, it will help you make decisions that are more true to the life you want to lead.
Key Insight: Did you know the frequency of the words you use is determined by an underlying mathematical pattern? This mathematical pattern is a power law known as Zipf’s Law. The underlying mechanism for Zipf’s law is not yet agreed on, but the main hypothesis is that it’s an outgrowth of the Principle of Least Effort. In short, systems that survive and operate at steady-state optimize for efficiency. When they do, things tend to look like Zipf distributions. Link
Unpacking stock-based compensation: Morgan Stanley has a comprehensive report on the nuances of stock-based compensation. This article's clarification that accounting adjustments modify how numbers are presented is one of the things that makes it useful.
Key Insight: Investors' perceptions of companies aren't much changed by moving stock pay from a footnote to the P&L or presenting an adjusted earnings metric that does not include them. Because there is a smaller population of investors that businesses can actively mislead through their disclosures, it may be advantageous in this situation that retail investors tend to be less sensitive to valuation. Link
EAR TO THE GROUND.
Podcast Episode: The Past, Present, and Future of Search
Guests: Sridhar Ramaswamy
Takeaway:
Sridhar Ramaswamy is the co-founder and CEO of Neeva and a Venture Partner at Greylock. In this conversation with Patrick O'Shaughnessy, Sridhar covers the history of search, what led to Google’s dominance, his view on the potential end-state for ad-based search engines, and how his experiences led him to found Neeva.
“Neeva is the world's first ads-free, private subscription search engine. [...] the logical conclusion of Google is that any commercial query that you put in is going to have nothing but ads, and any non-commercial query, a question about the last president, something about history, something about science, was going to be an answer that was provided straight by Google. It's a little bit of a dystopian vision for the future.”
- Sridhar Ramaswamy, CEO and Co-Founder of Neeva
4.15%
Apple introduced a savings account with a market-leading 4.15% interest rate, managed by Goldman Sachs. The new account needs an Apple Card and, for the time being, lacks a dedicated user interface. Link
PROFILE.
Alexandr Wang, the CEO of Scale AI, is profiled in Forbes. Scale specializes in locating people who can annotate data so that algorithms may use it. This does not scale well, and the business is primarily driven by price. However, Scale is also gathering some data for meta-models. When labor is inexpensive, one quality assurance method is to have two people annotate a given document and then confirm their findings when they disagree. Scale can partially automate the process of determining the ideal number of labelers. Enjoy the profile!
“As someone who is working on the most ambitious AI projects in the world, [...] I can tell you AI enhances and even supercharges humanity. AI always has, and always will, rely on humans. The long tail of real-world problems and the fact that there are always unknown unknowns means that humans will never be fully removed from the AI development lifecycle,” said Alexandr Wang, CEO of Scale.AI, at a TEDxBerkeley presentation in August 2022.
ON MY RADAR.
Codewand uses GPT-4 to build software with natural language. Link
About
Funston Brief is a newsletter by Funston Capital, LLC. We’re an investment company located in San Francisco, CA looking to acquire an already profitable and growing tech-enabled business. If you or anyone you know is interested in selling a business, please reach out to me at alex@funstoncap.com!
Thank you for reading. Like Funston Brief? Share it with a friend, or two. Have suggestions on how we can improve? Please connect.