Blue Meadow: What I learned from a year as a solo founder

Over the past 15 months, I took an unexpected detour in my life to work on a startup.

Because I spent most of that time as a solo founder, summarizing my thoughts in writing became my proxy for having a co-founder to bounce ideas off of. Similar to the rubber duck that a software engineer might be familiar with, writing is my way of forcing myself to examine what I’m doing.

I’m the kind of person who’s always looking at the next thing I want to do, and I miss a lot of insight as a result. It’s really tempting to put the last year of work behind me and move on, but I need to take a snapshot of my perspective first.

Here are some of the lessons I’ve learned, most of them the hard way.

Without meta-goals, entrepreneurship is a slippery slope

When I went back to MIT for my masters, I knew that I wanted to work on climate-tech. Although I was in a robotics lab and had sort of specialized in computer vision by this point, the problems I cared about were all environmental. Through a little bit of cognitive dissonance, I wrote a thesis and paper on uncertainty learning and stereo vision while planning my pivot towards something more meaningful.

I finished my degree during the first summer of COVID, and started looking for my way into climate-tech. At the time, it felt like there was very little for a roboticist to work on within the space. (In retrospect, there were many opportunities, and I just wasn’t looking hard enough). Then I learned about seaweed farming and was immediately hooked.

If scaled, seaweed can be a sustainable, drop-in replacement for terrestrial sources of biomass like corn and soy. There are a ton of cool startups making food, animal feed, plastics, fertilizers, biofuels, and more out of it.I became convinced that, like terrestrial agriculture, seaweed aquaculture needed automation to scale and become cost-competitive.

Thus began Blue Meadow, with the mission of automating large offshore seaweed farms with robots. To deal with some of the permitting and space constraints of near-shore aquaculture, we would even site the farms at decommissioned, offshore oil platforms.

I want to emphasize that I hadn’t even considered starting a company until this idea came along. My thinking at the time was something like: “COVID is happening... what a great opportunity for me to explore this idea I’m excited about.”

The hard questions I hadn’t asked myself were:

• What is the opportunity cost of starting a company?
• If this idea doesn’t work, will you find a new idea or go get a job?
• What are your criteria for giving up?

Since then, I changed ideas two more times and joined a blue-tech accelerator in between Blue Meadow v2 and v3. I went from ocean farming robots to an environmental monitoring system for seaweed and oyster farms, and then finally landed on a software platform that would streamline the carbon crediting process for environmental restoration projects with remote sensing.

I’ve discovered that I love the early ideation process of startups. When it came time to pivot, I always had more enthusiasm for the next idea than the one before. Since I hadn’t established a stopping criterion, I kept going and going, without questioning the opportunity cost of doing so.

Ultimately, opportunity cost caught up with me. Working on Blue Meadow had turned up so many cool climate-tech companies that I was no longer convinced that I would maximize my impact (or wellbeing) by continuing alone.

Be objective about founder-product fit

During Blue Meadow (v2), I was developing a sensor package that would be deployed on seaweed and oyster farms to collect environmental data. I should note that I have minimal hardware experience. I knew enough to start - I can do CAD, 3d print and laser cut parts, and connect sensors to a Raspberry Pi - but also knew that I would eventually be blocked by my lack of a hardware co-founder. That should’ve been an obvious red flag indicating poor founder-product fit.

I’m the kind of person who, for the most part, doesn’t let a lack of experience prevent them from diving into something new. But I’ve learned to (reluctantly) let an idea go if my skill set does not overlap with some core part of it.

Team > Idea > Hard Work

I had this ordering reversed when I started out.

Co-founders filter out personal noise

If I had to model my instantaneous level of optimism in pseudocode, it would be something like:

The rational signal is how “good” an idea actually is on paper, based on the data you have so far: feedback from customers, market size, progress on the product, etc. Added to that is some amount of personal bias. I tend to be biased toward optimism (i.e, that term is positive), although the bias term occasionally went negative after a few weeks of consistent setbacks or frustration in other areas of life.

Finally, there is a noise term, which includes things like day-to-day mood, how much sleep you got last night, reading a news article about your competitor, drinking too much coffee, spending hours trying to re-calibrate a 3D printer, or talking to someone who’s excited about your idea. All of these factors influence outlook, but are uncorrelated with how good an idea is; they are noise¹.

Assuming your co-founder’s life noise is independent from your own, then averaging together your levels of optimism should neutralize that term. I often wished that I had someone else in the loop to help pull me through the tough days, and keep me grounded when I was being naive or unrealistically optimistic.

Confidence is a practice (for me, at least)

Pitches, customer interviews, and networking events were all anxiety-inducing to me in the beginning. In general, I’m introverted and great at underselling myself; I was the CEO by default and not by choice.

All of these shortcomings improved with practice, but not permanently. If it had been a while since I’d pitched or done an interview, then I found the need to “warm up” again. My takeaway is that confidence can absolutely be learned, but if you’re like me, you have to keep using it or lose it.

Personal life downsides

• It’s hard to decouple your personality from your startup. For better or for worse, I became the seaweed guy.
• It becomes hard to justify technical exploration that doesn’t benefit the startup. When I was doing my masters, I still had the mental bandwidth for side projects (e.g building an indoor hydroponic lettuce garden), but not while working on Blue Meadow. That’s partly a personal choice that I made.
• You can’t train quite as hard. I trained for two marathons during 2021, and had a lot of nagging minor injuries that wouldn’t seem to go away. I also felt a little “flat” on more of my weekly runs than usual. I can't blame all of this on doing a startup, but I’ve come to appreciate the impact of stress (multiplied by 3-4 coffees per day) on physical recovery.

Collect enough information, and it all cancels out to zero

I first heard the above maxim on the Naval podcast, and encountered it a lot in practice. I’ll give a seaweed-related example.

There are two simultaneous ideologies within the emerging US kelp industry:

1. The industry should be made up of small, locally-run farms. Automation and consolidation ruined terrestrial agriculture, and are the enemy. GreenWave and Bren Smith’s book exemplify this point of view.
2. Seaweed farming is a promising climate solution that needs to be quickly and massively scaled. In order to reduce the $/ton production cost, we need selective breeding, automation, data-driven farming, and massive offshore sites that benefit from economies of scale. The ARPA-E MARINER program is a good example.

Although small farms and industrial seaweed production can coexist, these two directions have different sets of problems to solve and require different roadmaps. Many of the articles, podcasts, and companies that evangelize seaweed farming cherry-pick the advantages of both. I think it’s unfair to appeal to the environmental benefits of seaweed farming while opposing the technological advances needed for it to reach a meaningful scale (and price point).

These seemingly contradictory points of view confused me a lot early on. I started working on seaweed farming robots because of #2, and then got discouraged because I did the majority of my customer interviews with people in the #1 bucket.

Now that I think about it, Blue Meadow v2’s product (an IoT environmental monitoring system) was an attempt to satisfy both points of view. It didn’t automate a kelp farmer’s job, but it was a technological step towards improving farm productivity. I basically took the average of all the information and opinions I had collected and tried to turn that into a product (bad idea).

By choosing to solve one problem, you are necessarily choosing not to solve others. Being clear about which problems and opinions you’re not going to worry about is important for making progress.

Work on the problem(s) that give you the most FOMO

Towards the end of Blue Meadow (early December 2021), I started discovering all kinds of cool companies involved in carbon removal (e.g, NCX, sHYp, Phykos) and animal welfare (e.g, Transfarmation, and a multitude of alt-protein companies). I realized that if I didn’t contribute in one of those two domains, a huge amount of progress would pass me by over the next decade and I’d regret it.I’ve been using FOMO as a thought experiment to help myself figure out what I want to work on next. For example: “If I pursue problem X, and some other company solves problem Y, would I be disappointed?”

Time spent exploring an important problem is never wasted

The initial idea for Blue Meadow and the subsequent pivots were the result of open-ended exploration into problems that I thought were interesting or important. I’m certain that some of the obscure information I’ve dug up over the past year will become relevant in some unexpected way in the future. It always seems to.

Sources of optimism

1. How to Get Rich (The Naval Podcast)
2. Skin in the Game (Nassim Nicholas Taleb)
3. Any of Paul Graham’s essays