Coffee Shop Thoughts: 1/30/25 from Scottsdale!
I’m back at it with a few spare minutes, so I figured I’d take the moment to write. As luck would have it, I have a lot to share, so this will be a very shotgun post. The good news is that I have lived up to the name of this blog and the poem behind it.
Let’s go on an adventure Beyond the Yellow Woods!
Vet Clinic
I haven’t written about this yet but it’s time to let it out. My wife and I are building a vet clinic! Specifically, this will be an urgent care facility that focuses on providing emergency services.
It’s been a process getting here and lots of hard work with much more to come. However, we’ve been approved for a fairly substantial loan and just got our LOI approved on a building that we intend to purchase. As it stands, this will be a 4,000 sqft facility with a build out that encompasses 5-6 exam rooms, 1-2 surgical suites, and an ICU.
What’s more exciting is that this location is very close to our home and in a booming area with a desperate need for urgent care services. It’s a beautiful building too with floor-to-ceiling windows, a great view of the mountains, lots of natural light, and great access to amenities. Our goal is to open with 1 doctor but we already have verbal commitments from 3 other DVMs.
One of the primary reasons we decided to do this clinic was to generate a small but mighty counterforce to the private equity pillage and plundering happening in the industry. Our models have been designed and built around the concept of providing employees with profit-sharing and stock-option grants via RSUs.
In addition, we have a substantial budget to train veterinarian technicians into the equivalent of nurse practitioners. Today, most vet techs cap out at $25/hr and ultimately leave the industry due to a lack of career prospects. We believe that there are ways of increasing their pay to $40-$55/hr for top-tier talent and we intend to develop those individuals as such. Outside of that, we intend to set ourselves apart from the rest of the clinics by not only having a stellar culture (table stakes) but also a world-class benefits package. This goes beyond just the standard health and retirement benefits. I can’t say too much more on this topic but what I can say is that the DVMs and “high-performing individuals” in the clinic will have national access to some absurd “vacation” options. More to come soon ;)
This will be a longer build and will likely open in early 2026 but I plan on providing updates as I sporadically write here. Exciting times!
Starting at a New Company
In a totally random sequence of events, I came across two roads that diverged, and I took the one less traveled. Hopefully, it has made all the difference!
I was very happy with my role at Matthews. In fact, I’m more bullish than ever as to the work that we are doing. The CRE market is a $27T clusterfuck that needs to be disrupted. It’s plagued with insanely antiquated practices and absurd inefficiencies. A sad state all around.
Matthews has a vision of redefining what it means to be a brokerage firm and to broker physical assets. At the center of it all is revolutionary technology that will change the pace, resolution, and contextualization of the brokerage industry. As the boomers retire and sell off their assets, the millennial generation is stepping in. In the center of that is a significant gap between how CRE was transacted before digital natives and after digital natives. Matthews is solving this problem and has the mindset, growth, and positioning to do so.
I’m extremely happy with what we've built. So, why leave?
I was presented with a genuine “once-in-a-lifetime” opportunity to join a startup building some of the most unbelievable technology I've ever seen.
The company is P-1.ai
I cannot say much about what we are building publicly (yet). But, what I can say is that we are building a new foundation for artificial general engineering intelligence. I joined this company for 3 primary reasons:
To work with the most cracked af team being assembled. The folks who started and are joining this company are the 0.01% top brains in the world. It’s extremely rare to see such a concentration of talent and, selfishly, I’m eager to be the dumbest person in the room. I have high confidence that I will learn more over the next 12 months than I have in the last 6 years.
To usher in a new era of building the physical world around us. I strongly believe that LLMs will broadly disrupt and commoditize the majority of SaaS companies. Software as we know it is likely dead or fundamentally a “utility” moving forward. The biggest limiter to human acceleration is the bandwidth limitation between bits and atoms. To continue to go up the Kardeshev scale, we need significant innovation, leaps, and breakthroughs in how we build the physical world. P-1 is directly in the center of this.
To serve my country in an indirect way. While subtle, this is something that is philosophically becoming top of mind for me. We are in a race to AGI and ASI with the rest of the world. While the USA isn’t perfect, I sure as shit don’t trust the rest of the world with this technology. Outside of that, there are many threat vectors to war and to deterrence. I believe one of our biggest disadvantages right now is our ability to quickly build physical goods. This is outside of just assembling the “atoms” but also the system design of new things. We must be the far and ahead leader in deterrence-level capabilities in our arsenal and I believe that advantage starts with software that enables engineers to build in a new way.
I hope I’ll have more to say later on about this but I’m beyond excited, and a little bit terrified, of this new unbeaten path. This will be the hardest startup and role I’ve ever done but come hell or high water, I will help grow this company into one of the most successful startups out there.
The Race
There is a stupidly aggressive race right now going on. Foundation AI models. With the advent of LLMs, the pace of the foundation models is going to zero extremely fast. I firmly believe that LLMs will literally follow the same progressive utility slope as cloud storage. Things will just keep getting more and more cheap for the generalized foundation models just like storage did with things like S3.
That is the race that everyone is looking at right now. I believe there is a more important race that is starting to emerge. Ironically enough, the big data space had this right.
Volume, Velocity, Variety, Veracity, Value.
The 5 D’s of big data are ultimately what will matter the most in what will become the new frontier of specialized models: synthetic data. For specialized fields such as medicine, biology, or engineering where there are significant cross-domain dependencies (eg. thermodynamics, electromagnetism), our ability to not only collect a variety of data but generate scientifically accurate synthetic data to reduce training time via recursive learning will be the actual moat.
For example, let’s say you have a 5 component system within a variable environment. You know the metadata of each of the components but it’s impossible to simulate all of the interplay of these components within a variable environment space. If one component changes based on the environment, you have to simulate and verify the rest of the components and system are still viable.
Instead of computing all of the possible scenarios, you can synthetically train different scenarios of the environment which generates a wide parametric space. Think of this as the Dojo from the Matrix. The recursive learning nature of synthetic data significantly increases the accuracy of real-world scenarios while dramatically reducing the compute time.
Rest assured, it’s not as easy as it sounds. It requires a lot of upfront work on structuring the data appropriately such that the RL process works. However, once you get there, it’s a total cheat code to model performance.
If it were me, I’d be betting heavily on companies that are starting to go down this path. Just my 2c.
Short and brief this time around but hopefully some wisdom and insight for folks. I’m excited to share my new adventure with folks here in the next couple of weeks. In the meantime, I’m going to enjoy the rest of my trip here in sunny Scottsdale with a couple of margaritas and a great send-off with the stellar engineers at Matthews.
Onto a new adventure on the path less taken. Until next time!
- Ryan