2024: AI, Hackers, and the Rise of Pre-Seed Rounds
At Lombardstreet Ventures, we enjoy working with founders early on in their journey, usually during the pre-seed phase. There are multiple reasons behind our focus, but picking the top three, I would say:
a) it’s fun;
b) the investment decision is driven by how good the founding team is;
c) we believe that is where the best investment opportunities are.
2024 is a unique year for our company, mainly because the three things we are most excited about — AI, hackers, and pre-seed investments — are at the center of the AI revolution. We are still in the early days of AI, and we discover new opportunities and limitations every day. Compared to previous waves of innovation, like PCs or the Internet, which also emerged from Silicon Valley, this one is much different:
- AI is advancing rapidly, with significant breakthroughs occurring every few months.
- The big enterprise — or at least the brightest part of it — is catching up fast, adding AI abilities to its existing products, and this has been happening since day one.
- The primary stakeholder of the AI leading company, OpenAI, is Microsoft, an anomaly in a venture-backed company. Elad Gil wrote an article recently that provides the best explanation for this fact:
Cloud providers are easily the biggest funders of foundation models, not venture capitalists. Given they are constrained in M&A due to FTC actions, and the revenue that comes from cloud usage, it is rational for them to do so.
Some VCs think that AI will have a bigger impact than the Internet, and even though I don’t share that view, I’m enthusiastic to see an impressive number of founders creating and experimenting with thousands of new products thanks to AI. A strange thing I’ve noticed is that every time a disruptive wave of innovation is born, hardcore technical people — the hackers — start to flood the market again. They come out of their code caves and labs and gather to create companies en mass. This is precisely what is happening today, even if the hacker movement involved in this wave of innovation differs from those we were used to seeing in the past. They are not just nerds working from a garage night and day to build something cool, eager to release it online for free. This time, they are also well-educated college researchers or Ph.D. drop-outs obsessed with neural networks and Large Language Models. Reading online, I found out I’m not the only one having these thoughts—check this clip from Diane Hu, a YC Partner, during one of the last Lightcone episodes. Every revolution brings its craziness, and as has happened for every past innovation wave like crypto, new lavishly compensated roles are emerging. It looks like being an AI Engineer at OpenAI is a much easier path to getting rich than building a startup:
This [the AI engineer] will likely be the highest-demand engineering job of the decade [..] They are making $300k/yr doing prompt engineering at Anthropic and $900k building software at OpenAI. They are spending free weekends hacking on ideas at AGI House and sharing tips on /r/LocalLLaMA2.
—Latent Space AI article “The Rise of the AI Engineer”
Becoming an expert in AI technologies is challenging, but a million-dollar paycheck is a bit much.
It is worth noting that such anomalies are not uncommon in Silicon Valley. For example, in 2013, when DevOps and cloud infrastructure were gaining momentum, a well-known company in the Bay Area offered a $5 million package to an infrastructure monitoring expert my company was trying to hire. Of course, we couldn’t compete with that.
AI is hard. Even if I’ve been in the tech space for the last 30+ years, I must admit that I feel the struggle and keep asking myself: Should I understand this?
I can grasp most of the current AI-related innovations, but really understanding how they work is another thing. It is a fact — and a scary relief — that even the so-called experts do not possess comprehensive knowledge of the intricacies involved in the AI inference generation process. The reason behind it is that AI is not entirely predictable, which is part of this technology’s good and ugly, and that is probably one of the reasons why hallucinations are here to stay.
AI is not just a neural network researcher’s game; in fact, “when it comes to shipping AI products, you want engineers, not researchers,”—as the Latent Space AI article mentioned earlier. That is probably why San Francisco is at the very center of this unique movement; the city keeps attracting some of the best and most determined hardcore technical people on the planet, and things happen bottom up. As Jeremiah Owyang noted in a recent LinkedIn post:
SF Bay Area has more AI events in one single month than most countries will have in an entire year [..] 140 SF Bay area AI events in Feb 2024, many of these events are hackathons where developers build, workshops, showcases — not just talking about AI.
Nobody decided to build an AI community in San Francisco; it just happened, and people loved it. Hackers aim to stick together and possibly share common houses, “blurring the line between work and play”—as Liz Lindqwister from The San Francisco Standard wrote. Hayes Valley and its neighborhoods became what is today known as Cerebral Valley, and the story repeats itself once again.
… hacker homes are a nostalgic return to the tech world’s roots, reminding us of the Palo Alto start-up homes — known as accelerators or incubators — that once spawned tech giants like Facebook. And as the generative technology industry becomes an increasingly popular and lucrative industry in the Bay Area, these hacker communities are more than just another iteration of tech’s unique lifestyle — they’re seen as the hubs of AI’s future.
— Liz Lindqwister
A remarkable difference between the prior and current innovation waves is that APIs are now considered the backbone of modern-day software development. AI + APIs’ large availability also means that new products can be released overnight, and founders don’t need to build for months or years before approaching investors seeking funding for their ideas. All this translates into a huge number of fresh seed-stage companies on the market, and that’s why the pre-seed segment remained strong despite the considerable challenges of the last two years. Although investors have become more cautious in their capital deployment, this stage has seen the most vigorous increase in supply in years. According to the Carta Data Minute report by Carta Insight, pre-seed rounds have grown by more than 50% since 2021:
… total capital invested is down across nearly every venture stage if you compare 2023 activity to 2021. That’s what all the headlines have been blaring!
But when you segment the data more granularly, an interesting pattern emerges.
Fundraising for pre-seed companies (defined here as companies raising less than $1 million on a SAFE or a convertible note) is actually more than 50% higher than in 2021.
This is why I’m bullish about the next 24 months. Living in times of rapid change, technological advancements, and uncertainty can be both challenging and full of opportunities.
