Infopunk #6 – What real learning looks like
What real learning looks like, Deep Dive into Deep Tech & Julian Shapiro's take on life advice management
A startup I admire – This device lets you control your dreams | Prophetic
Prometheus stole fire from the gods, Prophetic will steal dreams from the prophets.
A song I like – I Wonder
What Real Learning Looks Like – 2-minute video. Just watch it.
Learner learning
Koko Xu – Deep Dive into Deep Tech
Koko is just brilliant and he happens to be a deep tech investor. He has developed a “theory of deep tech” that categorizes Deep Tech into 3 buckets:
· Single-Product Plays: Startups with highly technical or scientific risks, focused on one or a line of products. They have long time horizons, high defensibility, but limited scale potential. Cruise, the autonomous vehicle startup, could be enclosed within this category.
· Vertical Platforms: Startups creating a vertical solution in fragmented markets to provide top-notch services. They are characterized by long time horizons, high defensibility, and a decent return profile. Companies like Solugen fit these profile.
· Horizontal Platforms: Startups building infrastructure in frontier markets, characterized by short time horizons, high growth potential, and a great return profile. One of my favorite companies that fits this profile is Benchling.
Koko describes Deep Tech startups as "Tech-Centric Frontier Markets," where technological and scientific breakthroughs exploit opportunities in overlooked areas.
Some insights that caught my eye:
Science Fiction Products: One characteristic of deep tech is the development of what are termed as science fiction products, usually with some hardware element/hardware centric products. Deep tech often is about turning what was once considered fiction into reality.
Frontier Markets: Deep tech startups explore frontier markets, which are unchartered territories in the tech landscape.
Long Time Horizons: Especially seen in Single-Product Plays and Vertical Platforms, the long-time horizons represent the difficulty of the problem these companies are trying to solve. Counterintuitively, I think this “time arbitrage” is one of the advantages of Deep Tech.
Capital Intensity: Deep tech tends to require significant capital investment due to the high technical and operational complexities, and in many cases, a long-time between company founding and first revenue (due to a lot of expenses being carried out in the prototyping and initial product development phase). I think this is also a feature, not a bug, of deep tech investments.
Market Risks and Defensibility: The market risks are high but so is the defensibility, especially in Vertical Platforms. It can be easier, due to the difficulty and intractability of the endeavor, which deters potential competitors from entering the market, to create competitive advantage over time.
Macro Trends
Below you can find a few more non-obvious observations that are not necessarily technology-related:
Dual Research-Product Startups
Since the days of Xerox PARC, Bell Labs, and Los Alamos Laboratory, high-flying research organizations have been known as hotspots for innovation but flops when it comes to monetization.
Recently, startups like DeepMind, OpenAI, and CTRL-Labs have pioneered the playbook for successful Dual Research-Product organizations. The flywheel looks something like pursuing short-term high-margin partnerships, using the proceeds to fund cutting-edge research, building tooling and infrastructure around scientific breakthroughs, then rinse and repeat.
New Age Academia
Technology Transfer is the process of spinning out research from academia to industry. Well-funded universities have Technology Transfer Offices (TTOs) that provide a suite of services intended to help accelerate the process. They are generally shit at their job. TTOs around the world are known for their fickle terms and timelines, predatory negotiation tactics, and low translation efficiencies. TTOs rely on an unmotivated committee of volunteers to evaluate which Patent Disclosures are worthy of funding, and as a result, 97% of all patents filed through TTOs are never licensed.
Koko has spoken about this topic with Marc Singer, cofounder and Managing Partner of Osage University Partners - an $800m fund investing strictly in Tech Transfer startups; Teresa Fazio, who leads commercialization at MIT’s Lincoln Lab - the leading national R&D center; along with dozens of TTO operators at universities ranging from Stanford to NYU. The feedback is unanimous: despite the tremendous inefficiencies in Tech Transfer, there is still an overwhelming supply of high-quality applied research ready to be commercialized. The bottleneck is unequivocally entrepreneurial talent.
I suspect as more internet entrepreneurs seek their second acts in higher-impact fields, we’ll see this imbalance inch closer to equilibrium, resulting in more Deep Tech startups with defensible core technology and experienced entrepreneurial operators.
There are more grad students and postdocs than ever before, with fewer and fewer faculty positions available. The average age of professors is now 55 years old. As a result, many young academics are abandoning the siloed and low-paying academia bubble for collaborative and highly compensated research positions in industrial labs.
This talent arbitrage, combined with the fundamental research constraints of academia in categories like Artificial Intelligence and Robotics - where data, compute, and hardware are all easier to acquire through corporate-sponsored research than shoe-string grant budgets - is causing a brain-drain away from academia and increasingly into technology.
Where there is talent, there are returns. Initiatives like the Venture Science Doctorate offered by Deep Science Ventures are a testament to some of the ways in which this talent arbitrage is happening.
Academia is broken in many ways, from the reproducibility crisis to the citation system to the oligopoly of commercial journals. Much has been written about this field of “Metascience” in recent years.
Koko has interviewed nearly every expert in this field for his last startup - from Brian Nosek (founder of the Open Science Foundation) to Ben Reinhardt (founder of Spec Tech building PARPA) to Adam Marblestone (CEO of Convergent Research building FROs) - and built a hive mind of Metascience-related readings.
These individuals and organizations are innovating at the Valley of Death between academia and industry - I won’t go into more detail as all of the aforementioned individuals have written extensively about this topic, but the takeaway is that there is a class of emerging Independent Research Organizations where many Deep Tech startups will spin out of beginning around the year 2028.
A friend recently said to me that “Deep Tech is becoming a buzzword”. While this was definitely not the case a few years back, there are signals, to take one example, that the Defense market is forming a bubble. My chief concern for Deep Tech is an observation I’ve made that talent always follows capital flow, but a few years late.
If Deep Tech is becoming buzzy today, we should expect the next 3-5 years to be increasingly filled with noise - too many dollars chasing too few high-quality entrepreneurs. That said, as outlined in this primer, there are clearly many good businesses to be built in Deep Tech, not to mention its outsized impact on our world when compared to plain vanilla software.
For investors, the “Pick” part of the “See-Pick-Win-Help” cycle will play an outsized role in generating great returns in Deep Tech for vintage 2023 - 2025 funds. For founders, having an air-tight thesis using some of the frameworks above will put you above the crowd. In the immediate two years, huge funds raised during top-of-cycle (namely vintage 2021) are stuck with too much cash to deploy - they’re desperately looking to back bold, capital-intensive theses, partly for the huge upsides, and partly to just get rid of the money.
These startups usually live in big markets. The best time to build a Deep Tech startup was probably 2019 (cheap money without bad habits), the second best time to build a Deep Tech startup is today.
Julian Shapiro - Memorized Rules: How to give your life direction
It's funny how we often come across life advice that strikes a chord, and yet, before we know it, it's lost in the hustle of our daily routines. This article taps right into that experience, highlighting a couple of key reasons why we struggle to turn advice into action.
First off, there's this common misconception about life advice. We tend to treat it like a random fact we need to remember – like someone's name at a party – and then we're surprised when it slips our mind. But life advice isn't a trivia fact; it's more like a set of instructions or guidelines, similar to what we learn from textbooks. The real trick is to drill these principles into our daily life, thinking about how they apply to the problems we're facing right now.
Then there's the issue of advice overload. We live in a world brimming with wisdom, tips, and hacks. It's easy to get overwhelmed. That's where the idea of "Memorized Rules" comes in – it's a brilliant strategy to keep things manageable. By focusing on a select few rules – say, six, based on our cognitive limits as outlined in Miller's Law – we're more likely to remember and actually use them. This approach simplifies the overwhelming task of applying advice to our lives, making it practical and actionable.
What's really fascinating is how these Memorized Rules start to reshape who we are. They slowly seep into our identity, guiding our daily decisions and interactions. It's like they become the framework through which we view the world, influencing everything from our relationships to our goals.
The article also offers some wonderful examples of such rules, which is super helpful for anyone who's looking to adopt this method. It’s like a guide on where to start, showing us the kinds of principles that are worth embedding in our minds.
And let's not forget the importance of emotional and logical connection. For a rule to truly stick, it needs to resonate with us, both in terms of making logical sense and striking a chord emotionally. When a rule manages to do both, it’s far more likely to become a part of who we are and influence our behavior.
Our past experiences also play a huge role in how we process advice. It’s these experiences that provide a foundation for advice to really mean something to us. Without relevant experiences, advice can feel distant and hard to apply.
The distinction Julian makes between general life rules and values is quite insightful too. It's an important consideration that helps us prioritize the advice we focus on – separating general guidance from those rules that tap into our moral compass and emotional fulfillment.
And finally, the practical steps on how to integrate these rules into our lives – that’s the real game changer. It’s not just about picking the rules but also about revisiting them, adapting them as we evolve, and ensuring they continue to resonate with us.
See you on the next issue.
Stay curious,
Chief Infopunk