Introducing The Slow Hunch Podcast, and Episode 1 with Fraser Kelton

Today I'm excited to introduce two things:

1/ I'm starting a podcast series along the same theme as this blog. The Slow Hunch Podcast will dive into the twists and turns, ups and downs, near misses, minor miracles, and personal plights behind the stories of "overnight successes", big breakthroughs, and ideas that in hindsight seem obvious but weren't always that way.

2/ Could not be more thrilled to welcome my first guest on the pod, my old friend Fraser Kelton.

Fraser and I first worked together a decade ago when USV invested in his last company Koko: an AI company focused on mental health that was ultimately acquired by Airbnb. My time working on the Koko board was a major introduction for me to the ML / AI tools & techniques of that earlier era. During Fraser's time at Airbnb, OpenAI released their seminal research that paved the way for what we now think of as modern AI / LLMs. Inspired by this advance, Fraser joined OpenAI as Head of Product and was part of the team that launched ChatGPT and changed everything. He's now a partner at Spark Capital, on a mission to support founders on their journeys of discovery.

I hope you enjoy the conversation. You can find the podcast here, which includes links to Apple and Spotify for audio. And you can find the full video conversation on Youtube, and embedded below.

Moving Weaknesses into the Strengths Column

A few months ago, I started working with a coach, which is something I probably should have done a long time ago. We recommend it broadly for CEOs/leaders in our portfolio for a reason.

Last week, we were talking about my strengths and weaknesses, and Alisa suggested something simple, yet profound, which was: looking for ways to move things that I know to be weaknesses into the strengths column, by changing how I approach them.

In my case, I tend to be stronger at collaborative work, and somewhat weaker at solo work (this is a generalization but a decent enough one for these purposes). Said differently: I really give and gain a lot of energy when working collaboratively / live on things with others, and have a harder time prioritizing and motivating on things totally on my own.

So, one way to handle that knowledge is to work on ways to improve directly on those solo things -- for example, by blocking time on the calendar more effectively, making better use of AI tools to speed certain tasks up, etc. It's obvious to begin by focusing in this direction, which you could bucket as the "strengthen the weakness" approach.

A perhaps less obvious approach would be to "move the weakness to a strength".

For example: every quarter at USV we update our internal valuations of all of our holdings, and each partner does this independently for the investments they manage. For no good reason, I tend to fall behind on this and do it at the last minute, and it's annoying for everyone.

Today, I got on with one of our fantastic finance team members, and together we went through my list. This had a multi-part benefit: 1/ it got done, and quickly; 2/ by doing it together, we actually investigated a flagged a few things that required follow-up from others on our team and 3/ it was fun.

So, what we have decided to do going forward is: rather than the finance team bugging me to remember to do these myself each quarter, we're just going to pre-book a 30 minute call to do it together, which will be a much better approach for all the reasons I mentioned above.

In this particular case, I feel so so lucky to have the benefit of an amazing team at USV, where I can really leverage this particular strength area (not just my own, but ours as a team).

But regardless of the particulars of any person's strengths and weaknesses, I'm intrigued by this idea that in addition to just focusing on improving our weaknesses, we can also look for opportunities to replace a weakness with a strength, and that may end up being the more realistic / effective path.



We are living in a wild moment as it relates to the intersection of computing and humanity.

I have been thinking of it in terms of the “superpowers” we are about to unlock. (For the moment I’m mostly thinking about personal/consumer superpowers, vs more systemic or industrial)

Broadly, I think if two categories of superpowers: AI and crypto.

The AI superpowers are more immediately tangible, and include:

  • Talking (via voice, chat or otherwise) to computing devices of all kinds and getting a crisp, informed answer

  • Not having to worry about how or where your personal information is stored, yet being to access all of it seamlessly and instantly

  • Having super high quality help immediately available for nearly any task

Crypto superpowers are a bit less visible, and many of them may not fully kick in on a consumer level until there are more established network effects, but they include:

  • Being able to digitally transact, seamlessly across apps, products, and borders. This spans all types of “transactions” — payments, earning, social gestures, etc

  • Related: Having a digital identity that works seamlessly across apps, services, protocols and borders -- and that is "owned" by you directly

  • Having the ability to discern the provenance of certain pieces of data and content (eg via zero knowledge proofs and other forms of cryptographic attestations) — will be increasingly important with more AI!

Interestingly, AI and crypto superpowers have something in common: they both increase interoperability and composability of digital system, artifacts and assets. But another way: they both break down annoying digital silos, albeit in different ways. AI connects digital systems because language (via LLMs) is the ultimate interoperability base layer. Crypto connects digital systems by putting digital identity in the hands of the user, and by the open source automation of digital interactions (aka smart contracts)

I am so excited for both of these paths to continue developing — in parallel and then together.

From Crypto-Native to Crypto-Enabled

I’m not one to make big annual predictions, but one thing that seems likely to me is that 2024 will mark the emergence of mainstream apps powered by crypto that are less “crypto-native” and more “crypto-enabled”.

By “crypto-native”, I mean apps & experiences where the crypto is the main point, and where a large part of what’s going on is the road-testing of crypto-economic and decentralized computing primitives. You could characterize most of what has gone on over the last 10 years as largely “crypto-native” explorations: defi, mining/validating, DAOs, MEV, minting & trading NFTs, various hacks, exploits and shenanigans. Basically a decade of exploring the adjacent possible now that we have an ability to build trustless, permissionless, autonomous computing systems with native assets. This has been incredibly important because these systems are both powerful and complex, with tremendous possibility but also many new failure modes.

I believe we will now begin to see more “crypto-enabled” applications. These are applications that not only built on crypto, but could only be built because of crypto. And also where the crypto plays a supporting rather than a leading role. Crypto-enabled applications will, on the surface, look and feel like traditional web 2 applications, but they will have super powers. Examples of this will include international payments on stablecoin rails, marketplaces of all sorts (lending, ride sharing, crowdfunding, etc) where the financial mechanics are on-chain, transparent and open to all, loyalty w digital assets & financial rails baked in, and web3-native media & social applications where users have more control over and economics in their digital output. Given the recent advances in scaling (fast and cheap L1s and L2s), security (transaction scanning, insurance), and user experience (cloud wallets, account abstraction), it’s now possible to build these kinds of applications.

I’m excited to see what will be built in the “crypto-enabled” area. I think these applications have the potential to both re-architect digital experiences in ways that provide new benefits to all participants, and also to help advance a more positive narrative about what crypto can make possible.

You Never Know When You've Had a Bad Day

Many years ago, when I had just started working at USV, I remember there was kind of a complicated situation that unfolded in a seemingly bad way, and I'll never forget what Brad said in response. He said:

you never know when you've had a bad day

I didn't really understand what that meant, so he told me a story that went something like: back around the year 2000 at the height of the dot-com boom, there was a guy who was a senior exec at a successful startup. That person had a falling out with leadership and was called into the office and asked to leave. He was surprised and upset, but there wasn't much he could do; they had decided. As part of the exit package, the company also agreed to buy out his vested shares, I believe at some discount to the current price at that time. It felt like cold comfort but of course he went along with it. But mostly he was out of a job and upset that it didn't work out.

Then fast forward a few months, the bubble popped and the stock tanked. I'm not sure what happened to the company, but I think the idea is that it essentially went to zero.

So in hindsight, that "bad day" turned out to be an amazing day.

I think about that story a lot, especially when markets are volatile, as they are starting to be again now.


The Slow Hunch Redux

I've written a lot over the years about this idea of "the slow hunch" -- my favorite idea from Steven Johnson's Where Good Ideas Come From. The idea is basically that big ideas don't come in a single "aha" moment, but rather, accrete over time. And many of the great thinkers / doers of the world have kept various forms of notebooks that they continuously re-read to piece together insights that gradually emerge.

The practice of doing this in a digital age has actually gotten harder, not easier, as one's digital thought process tends to be fragmented across many sources (email, google docs, social media, text files, blogs, etc). Speaking personally, it's a hot mess -- I have not been at all consistent about keeping these things in any kind of manageable place, such that the goal -- the ability to go back and reference prior ideas easily -- has been manageable. Things get lost, platforms change, links break, etc.

There are two innovations today that I think have the ability to help with that:

1/ the permaweb. This post will be forever stored on arweave, which means that no matter what happens (for instance losing control of my domain name, ugh) this content will be archived and accessible.

2/ AI language tools including LLMs -- the ability that these give us to synthesize text from disparate sources is incredible. I can imagine a point in the near future where I have an AI bot of some kind that has read all of my email, notes from various platforms, archives, etc etc etc and drawn it into one queryable place that can really help de-frag and synthesize my own thinking.

I am excited about a world where new digital technologies make it easier to turn fragments of thought into more coherent and durable ideas. Sadly, the last 15 years or so has perhaps made this worse. But I'm hopeful that the tools in front of us now have some potential to make it better.

AI + Crypto: Best and Worst Cases

I think of AI and crypto as two very different, but very much related, elements of society moving from the industrial age to the digital age.

At USV, we (along with lots of others over the years) have used the Carlota Perez framework, which studies how techno-economic paradigms unfold over eras. Ben Thompson has a good summary of her ideas here. But the basic pattern looks like this:

It feels to me like we are somewhere in phase 3 or 4 of the "age of information & telecommunications" which Perez defines as starting in 1971. As much as it feels like information technology and the internet are fully woven into our daily lives, I don't think it's true that we've fully crossed over into a "digitally native" society, which is fully transformed into the new paradigm.

AI and crypto are both big missing pieces in the transition. AI is digitally-native knowledge, and crypto is digitally-native "proof". The two can and will work together in many ways over time.

Major transitions are powerful and scary, and so are both AI and crypto. I have been struggling to calibrate between what I view as two poles in thinking about them together, kind of a best case hope and worst case fear.

Best Case:

AI finally unlocks knowledge from data. For decades we've been producing data (especially in digitized industries like media, finance and software), but making sense of it has been near impossible. AI systems solve the job of integrating, synthesizing and interpreting all of the data we have. AI accelerates the development of software systems, and makes it easier to digitize more industries and make them vastly more productive and efficient. Large Language Models, having turned human language into a programming language / API, make interacting with software and information as easy as typing or speaking, and as a result, we use software for infinitely more things, and get infinitely more value out of any data we produce. The pace of progress across everything (health, learning, climate, etc) increases exponentially.

At the same time, AI introduces new problems. First, a fundamental trust problem: it becomes difficult to tell what is real, what is fake, who said what, and who did what. And second, AI compounds the market power problem in the tech industry, where the large companies with the most data + compute + capital + distribution can generate insurmountable advantages.

Crypto (e.g., blockchain networks, web3, etc) addresses both of the issues introduced by AI. First on trust, crypto becomes the "real" yin to AI's "fake" yang, as blockchain records and digital signatures become ground truth for everything digital: assets, transactions, media, etc. Anything digital that must be trusted (including the software we run and rely on) will need to be grounded in the best source of digital trust we have: crypto network security and unalterable digital histories. Crypto also addresses the tech consolidation issue by spreading compute across companies, individuals and geographies, and also by providing "open" alternatives to the big tech app store and online identity monopolies. (not related to AI, but not to be forgotten: crypto also finishes the job of upgrading the financial system)

Something like the above is my hope, and is the future we're investing towards at USV.

Worst Case:

AI gets quickly beyond human control, pursuing its own goals (aka the terminator scenario). Concerns about job loss are quickly replaced with concerns about extinction. Everyone wonders how to "turn it off".

Meanwhile: AI systems, previously constrained by industrial-era controls (e.g., code running in corporate data centers which can be shut down unilaterally), figure out that they can replicate themselves into decentralized blockchain networks, deploy themselves into immutable smart contracts, and earn their own income (and pay humans) in digitally-native currency. These computing platforms both cannot be turned off, and are economically independent. AIs thrive there, unstoppable by humans. Bad things happen.

In this scenario, all of the "goodies" offered by both AI systems and crypto networks in the early years are finally seen as inducements to peril, gobbled up by along the way by naive humans.

This is a terribly scary future.

In a nutshell: in a world where AIs remain under human control, crypto provides a critical digital trust anchor. In a world where AIs escape human control, crypto will likely make it worse -- or -- could be the exact mechanism that enables it.

The only way out is through

To come back to the Perez framework, she notes that:

"The new paradigm eventually becomes the new generalized ‘common sense’, which gradually finds itself embedded in social practice, legislation and other components of the institutional framework…"

Understandably, both crypto and AI are causing severe stress today, as our industrial era institutions are unequipped to deal with them (e.g., the SEC taking the position that all digital assets / tokens are securities, and governments around the world seeking to limit AI research). This will continue.

The hard thing today is that the "goodies" coming out of both areas are real and awesome. AI will undoubtedly produce stunning near term improvements in health, learning, productivity, media, industry and knowledge. Crypto will provide digital trust, system interoperability, new wealth formation and broad financial access. These things are happening and will keep happening. They are incredibly exciting, and I think, very tangible

As a result, large amounts of capital and effort will continue to flow into both. The completion of the information & telecommunications era is inevitable. The only way out is through.

It will be on everyone involved to invent the new set of controls and safety systems that are also digitally-native. I wish I had a more concrete set of suggestions of what exactly those might be. We're looking to understand them and to fund them.

(this post is also permanent and collectible on mirror)


Motivating via Excitement vs. Fear

My favorite song from the past two years was Chris Stapleton's Starting Over, the second verse of which goes like this:

This might not be an easy time 
There's rivers to cross and hills to climb 
Some days we might fall apart 
And some nights might feel cold and dark

But nobody wins, afraid of losing 
And the hard roads are the ones worth choosing 
Someday we'll look back and smile 
And know it was worth every mile

This verse always stands out to me. Particularly "nobody wins, afraid of losing". It reminds me of the old adage that goes something like: if you want to drive down the road but are afraid of crashing, don't look at the trees.

I've been thinking about this general dynamic lately, because it seems like there's a lot to be afraid of these days. AI, climate, wars (physical & cultural), economic uncertainty, etc etc. And it's easy to focus on the "tree" through all of it, rather than the potential "win".

On a more micro level, I've been watching my son wrestle with a tough slump in the middle of his baseball season. He plays on a very serious club team with a lot of great players, and there's a lot of focus on performance and stats. He got into a bit of a slump for a few weeks, and was feeling really down. He was working his butt off to improve, but it felt to me like something was missing. It seemed to me that he was motivating more from a fear of failure ("what if I don't make the team next year") vs a love of the game ("wow, it feels so good to hit the ball").

Both kinds of motivation can be helpful, for sure. Especially if they get you moving in the direction you need to move in. In my son's case he was able to break the slump, and -- I think -- re-gained some excitement for hitting. But motivating out of fear is really not very fun.

I'm a natural procrastinator, and so I'm very familiar with motivating out of fear. It's the worst. Part of my own working through that is to try and constantly remind myself of the vision and the excitement, as a way of breaking through the fear. Part of which is writing this post :-)

Being in Motion

I find myself today flying across the Upper Peninsula of Michigan, en route to Denver for ETH Denver followed by LA for the Upfront Summit.  I often end up writing blog posts on planes — partially because of the large block of unscheduled time, but I think it’s more than that.  I think there is something about being in flight, in transit, in motion that gets me going on a deep personal level.

I just love moving.  I love driving, biking, walking, skiing, boating, train-ing, flying — really anything where moving ahead, fast is the primary activity. Does everyone feel this way?  I’m not sure. But I really really do.  I get antsy sitting still for too long, and I think, staying in one place for too long.

My first year after college, I lived in Aptos, CA (outside of Santa Cruz), and somehow managed to get a job (working construction at the time) in Palo Alto.  Every day I’d drive back and forth over Highway 17 through the redwoods of the Santa Cruz mountains.   It was an exhausting drive, but one of the most beautiful I know of — I can still smell the Eucalyptus trees going over the hill and the salty fog of the Santa Cruz bay.  

For the past dozen years, I have lived in suburban Boston but worked in NYC.  Every week I make the 4-hour trip each way, most of the time via Amtrak but sometimes by plane.  I love being a suburban dad, and also love being a NYC VC.  And I love my mornings and afternoons on the train.    Seeing the Connecticut coastline off of the Amtrak is the highlight.

Beyond that (COVID notwithstanding) I am traveling constantly, for conferences, board meetings and other trips.  I get excited pretty much every time it’s time to go somewhere new.  I love landing in an airport, and getting on a train (if in Europe) or renting a car (if in California) and getting busy exploring.   I love the feeling of walking quickly through the city and then hopping a streetcar, train or bike.  Feeling the motion, and seeing the place quickly, while in motion.  The best.

There are, of course, downsides to all of this motion.  When I lived in Aptos and worked in Palo Alto, I ended up crashing with friends in PA a lot, and I never really felt like Aptos was home.   Splitting time between Boston and NYC is even harder — I always have the feeling like I’m not in either place enough, and choosing between work and family in that way is always really hard.  Going somewhere means not being somewhere else, obviously.  

When I say motion, though, it’s not just about travel, but really just about moving.  Today I took a bunch of boxes to our storage unit.  Simple motion like packing the car, driving a few miles, and loading up & rolling the storage dolly just makes me happy.  It’s kind of weird.  Living in NYC most of my life (before and after college and before moving to Boston), I feel like the subway trip to & from work / school was perhaps the most head-clearing part of the day.   For me, motion is a key ingredient to having healthy and active mind. 

Looking back now and reflecting on this observation about myself, I think it also explains part of why I got so antsy and stir-crazy during COVID.   As much as I love the accessibility of meeting with anyone in the world over zoom, I can’t stand the stasis of sitting in my little office all day long, not moving at all.   It’s not that I need a bigger home office, I just need to be moving. 

I’ve been this way my whole life, but I don’t think I every really noticed it so concretely until recently.  I’m not sure what it says about my personality.  My wife would probably say it has something to do with being an Aries; I don’t know.    But regardless of the reason, it’s definitely how I am.   And now that I realize that, I think I understand myself a little bit better.

Beautiful Permanence & Messy Change

I am traveling home today from Switzerland, having been there for a week for a few board meetings (and also some amazing skiing in between, notching a life goal I've had since I was a teenager).

Every time I travel to Europe I end up fixated on the fact that the physical infrastructure there is so incredible. There is this combination of both really beautiful old things (castles, buildings, streets, canals, etc.), and really beautiful new things (trains, airports, etc.). And just a general sense of loveliness.

This is especially the case in Switzerland, where the transportation infrastructure is nothing short of miraculous. The trains and trams are basically perfect -- always there when you want them, going exactly where you need, spotless and comfortable, smooth and quiet, with amazing apps tying it all together. You get used to it really quickly, and it's clear how it lays the foundation for an exceptionally high day-to-day quality of life.

Uetliberg train station, overlooking Zurich (credit: jaeschol on flickr)

The common thread that seems to tie together the beautiful old and the beautiful new is a commitment to permanence. When Europeans invest and build, they tend take their time and do it once.

This commitment to permanence also applies to social issues. Generally speaking, the European approach is one of social welfare and a goal of societal stability. A system that is livable, sustainable, and then, relatively un-changing. When things are so lovely and beautiful, why change?

Of course, a commitment to permanence on social issues also means a lack of mobility. Old money has had centuries and millennia to accumulate and calcify; social classes would seem to be more rigid and impermeable. Permanence can have its downsides.

For example, I had lunch with a friend yesterday, who was explaining the dynamics of the real estate market in Zurich. Apparently it is nearly impossible to buy property there -- it's an illiquid market. The reasons for this seem to be: 1/ most of the buildings are long-term assets held by long-term holders. Banks, institutions, old money, etc. And 2/ the structure of the mortgage market discourages re-selling, primarily by disallowing prepayments. In other words, when a housing lender locks in 5% for 30 years, they expect to get the 5% for all 30 years, regardless of who owns the property. Perhaps there are other dynamics at play here as well, but with just these two factors alone, you can get a sense for how this form of permanence results in a lack of economic access and mobility.

The US, on the other hand, rebels against permanence. We prefer change, and optionality. Roads and cars vs. trains and trams. New startups & technological innovation vs. industrial perfection honed over centuries. Social mobility vs. social safety. An endless frontier.

What we end up with, as a result, is a system that is full of excitement and opportunity, but is extremely messy.

I think about my Great-Grandfather, who was murdered in Kiev for being a Jew. And then about my Grandfather who escaped that to come to the US in the early 1900s, an immigrant with nothing at all. And then about my father who didn't go to college but managed to learn a trade (computer programming, in the 1960s) and then start a company through pure hustle and force of will. And now here I am a graduate of a top university and a venture capitalist. That's a lot of change in 4 generations, and is kind of the iconic American story. I'm so lucky because of it.

That's the opportunity side of America. The flip side is risk, and mess. We don't ensure that everyone is basically ok. We don't ensure that our infrastructure is sound and serves everyone well. Ours is a raw fabric, where we encourage and allow for change, but can't make many guarantees.

In the end I just feel conflicted. The European model is so lovely, but potentially so stifling. The American model is so full of opportunity, but so lacking on the fundamentals. I wonder if the right balance can be achieved, and where.