There’s no better way to show you have high conviction in yourself as an investor than being the biggest LP in your $101 million fund, right?
Especially if you name your firm Conviction, as Sarah Guo did after leaving Greylock following a decade of investing for the well-known venture group. Last week, she announced that she raised $101 million for her new fund to back companies that are building artificial intelligence and what she describes as software 3.0.
Guo spoke to TechCrunch’s Equity podcast, co-hosted by Natasha Mascarenhas and Alex Wilhelm, about her inaugural fund and the broader market that she is investing in today. The entire conversation is live now wherever you find podcasts, so take a listen if you haven’t yet. Below we extracted four key excerpts from the interview to discuss further. Guo’s comments were edited lightly for clarity.
Think of venture in innings
Part of the allure of startups is that when things don’t go wrong, which they often do, you might just find yourself as an early employee of a rocket ship. That counts in VC, too, of course, if you were the first person to back a company like Airtable or see the power of connected fitness.
But what happens when you want to disrupt a category that has been around the block a few times? Guo shared her framework around venture innings and how that plays a role in her new focus areas at Conviction:
I was not the person to discover SaaS, right? I learned a few things about what made great companies uniquely good, but I kind of caught the shift to cloud midstream when I joined Greylock in 2013.
I got to learn something from really pioneering SaaS companies: Workday, Figma, Quip, Dropbox, Okta, and then a bunch of cloud infrastructure and security companies. But you know, we’re not in inning two. Are we in inning seven or inning nine of the cloud transition? And if you’re an early-stage investor, you definitely don’t want to be investing in inning nine in general. I’m all ears to important new SaaS companies that just figure out a better workflow or an underserved customer, but it is a red ocean out there now.
The primary clouds have been mostly transitioned and it’s going to get more expensive to go compete.
The thing that I see now is that you have this fundamental technology shift. The confluence of data collection, better data infrastructure, which has been happening for like five to 10 years now; to use the term from the from the academic paper, the unreasonable effectiveness of deep learning overall, and then like specific algorithms and then just being on the right side of growing compute capacity and deflationary generic compute, cheaper and cheaper compute.
I think that we’re really early in a period of both value creation and value harvest for more intelligence in software, like software 3.0. So what a time to be alive, like you gotta go after it.
Sensationalism AI and her framework for good companies
Many artificial intelligence-related companies tend to overdo the marketing and overhype the impact of their platforms. As we get to on the podcast, the hype has given users trust issues when it comes to AI. Guo noted:
I have a really strong allergic reaction to aggressive AI marketing, actually, because I think there’s been a long period of five-plus years of snake oil salesmanship around AI as a solution to all customer problems. It’s people overpromising and underdelivering, [but] there’s a lot of engineering work, and there are a lot of product hypotheses where you don’t know how well something will work until you actually work with the data in a customer environment.
Sensationalist AI aside, Guo’s framework for interesting applications in this space is helpful when trying to divvy up what she is and isn’t interested in. Below, you’ll see where she’s bullish right now, plus the opportunity for novel applications.
I think you can take a very clear-eyed view to the landscape and say, what’s valuable to a customer? I think there’s one way go sort of bottoms-up and be modality by modality, right? We can classify things. We can generate code. We can do math. We can generate images. And I think that’s an interesting one. [But] I think the way I tend to look at the world is to be interested in a set of problem domains that I know well because I know the customer well.
Nails, not hammers first. So you will see me invest in security infrastructure, developer tools, productivity applications, creative apps, generally enterprise-like sort of relational database applications that keep records [and] verticals where I think the vertical is large, interesting and the data is affected by this, like comp bio. The reason I think that software 3.0 is a really apt term is I’m just naming certain categories of software that I know well, but I fail to see a future where all of those [categories], given the advances in compute and data and algorithms, don’t get more intelligent.
I think that there are going to be completely novel applications of AI that don’t fit well into the existing categories. Visual generation is not an existing software category. Autonomy is not a software category that exists without AI. So I think there are going to be net new application categories … but I’m following the customer more than anything else.
Generalized AI is far away, sadly
During our conversation about AI and its applications, and measuring hype against reality, we asked about different models merging. In essence, if an AI product worked in one particular category, and another worked in a related space, could they combine in the future? And then could the combinations themselves combine? Here’s what Guo had to say:
You actually need to change product experiences to leverage the data [with] machine learning. And so I think the idea that there would be one ‘super app’ someday, because there’s one model that gets smarter from all of these different use cases, [is] pretty far away from that in terms of deployability.
Parsing the investor’s perspective here, our idea is a no. It appears that AI models that are useful for one particular thing or another will remain somewhat siloed. At least so long as we are tuning product experiences, to quote Guo, around what machine learning can teach us.
From this, we can infer that the advances we are seeing in bringing advanced AI/ML tooling to more products doesn’t mean that we are closer to any sort of artificial general intelligence, or AGI. The segment work isn’t summable to something greater than merely the collection of parts.
That doesn’t mean that AGI will never come; it’s more that it doesn’t seem like we have the stepping stones laid in front of us to get from here to there, based on our current work and investment. Bummer, but better to know than to hope beyond merit.
Why today’s startups are more fragile than we might think
There are a few competing trends in the startup game, including that it is at once cheaper and also more expensive than ever to start one. Sure, you don’t have to buy servers or pay for an office, but once you try to hire scarce engineering talent, you’ll discover that the “savings” that you racked up elsewhere gets very thin.
Not that startups have had to worry about cash in recent years. Venture funding was snowing from the rafters for a few years there, meaning that even the high cost of specialized labor was well within reach. Hell, startups used to tell us that their biggest hurdle was their inability to hire as quickly as they wanted to.
The resulting startups from that particular time-gated milieu are not robust. A bit too coddled perhaps? Here’s what Guo had to say during our conversation about changing venture requirements regarding growth and burn:
I think that the vast majority of startup companies today are just less robust than they were five years ago because they have had [access to] so much capital along the way. It’s counterintuitive, but you see more companies are burning cash when they go public. They’re not as profitable. [ … ]
But if you have this situation where cash is just fundamentally more expensive, then, yes, capital efficiency is going to be more valued. I don’t remember if I said this when I was on Equity before, but a deep belief for me is [that] I think an unhealthy habit of the last few years has been to think about venture fundraising as this unlimited pool of milestones that you need to reach to keep [your] company going. But venture capitalists are not the final authority. Eventually, companies are accountable to the public markets and their own profitability. Impressively, there are a lot of private companies who have a huge hoard of cash right now. And yet you can burn through $50, $100, $200 million in cash as a private company. It’s actually not that hard if you’re over-hiring and overpaying a lot of very expensive tech employees.
If you’re public or soon to be public, you’ll get the feedback that you need to become a more profitable organization quickly, right? But these [startup] companies with a big cash hoard, they’re often ending in a standoff with their boards [because] they don’t feel the pressure to be leaner while they’re wealthy, and eventually they’ll face the consequences.
If I think about my friends [who were founders during the dot-com era], they have no confusion about the importance of capital efficiency. And so I think you are, the current pain aside, going to get a generation of entrepreneurs that values capital efficiency deeply not because some venture capitalist or Natasha or Alex told them to, but because they have experienced that, you know, ‘I’m on a mission to build this amazing product for my customers, the outcome is the company’s profitable cash machine, and the less money I spend to get there, the better.’ That just wasn’t in the general culture of Silicon Valley over the last couple years.
Respect for the dollar, it appears, helps breed founders that care about not setting fire to too many, and the result of that work is companies that consume less to get as far — or farther, we’d argue — than those with access to unlimited funds.
In a sense, Guo is applying old-school venture wisdom to the current period; it’s an old concept that constraints breed creativity, and that too much capital can lead to startups losing focus. All that held up, it appears.
But with capital so cheap and available in the last few years, discipline itself lost its appeal, and we wound up with so very many startups that may still be cash rich but lacking in ability to focus or approach problems while on a budget.
Such companies are more fragile than their scrappier cousins. You might even call them less robust.