Aqua drone. Flood. Oh Water Drone Company. H2 air. Drone like a fish. Whatever I called it, it was going to be big. Gigantic. We probably will.
It was the pitch for my new startup, a company that promised to deliver one of the world’s most popular resources in the most advanced way imaginable: an on-demand drone bottled water delivery service. In my mind I was already picking out my Gulfstream private jet, shaking fists with Apple’s Tim Cook and staging hostile takeovers from Twitter. I just had to convince a group of venture capitalists that I (and they) were on to something good.
There were three VCs in total. The good news was that at least one of them was already excited about my idea. But he still had a few clues. I should, they suggested, focus on a niche market, be it athletes, office workers, or music festival-goers. They also wanted me to better explain why people should buy water from a drone instead of just picking it up at the store. Fair enough I suppose. Not everyone can imagine the widespread appeal of bottled water from the sky.
The second VC was a bit more on the fence. They had questions about the margins and the value proposition.
But the third was the killer. They told me in no uncertain terms that my unicorn wannabe was a “terrible idea”. Didn’t I know, were you wondering, that most drones have trouble carrying significant weight and that bottled water can be heavy? Even big players like Amazon are already focusing on drone deliveries and want to deliver more than just thirst-quenching water. Finally, the third VC broke out the worst news yet. “People don’t want their water delivered by drones,” they said. “It would be creepy.”
Criticism by robots
Whether it’s via email, phone, or in person, a variation of this conversation happens every day. Technology startups around the world raised a record $621 billion in venture capital last year, according to business analytics firm CB Insights.
However, the venture capitalists I communicated with were a little different. First of all, they had no money to invest. They didn’t exist either, at least not outside the ones and zeros of the software world. The three VCs – one positive, one ambivalent, one “kind of idiotic” – are the work of 44-year-old entrepreneur and anthropologist Francis Jervis.
Collectively referred to as PitchExpert, they are an attempt to automate domain-specific VC expertise and advice using the powerful GPT-3 AI language model. Originally released by OpenAI in 2020, GPT-3 is a massive, autoregressive language model that uses deep learning to produce human-like text.
Jervis previously used it in a product offered by his startup Augrented, which leveraged GPT-3’s wordsmiths to help people struggling to pay rent write automated letters telling them with their Negotiate settlements with landlords. Augrented users only had to write a few rough bullet points about why a discount was needed, and Jervis’ tool then quickly rendered formal-sounding, persuasive lines of text that could be added to a letter.
PitchExpert is based on a similar paradigm – except this time it allows founders to enter details about their potential startup company and receive a critique of it, including actionable feedback.
Long Tail Investing
This level of knowledge is a crucial factor in being a successful venture capitalist. VCs make so-called long-tail investing based on the idea that a small number of wildly successful investments (e.g., Facebook or Google) will more than offset a much larger number of failed investments elsewhere. It is a high-risk form of investing that is somewhat closer to gambling than traditional investing. Contrary to the typical bell-shaped distribution seen in stock market investing, venture capital gains appear more like erratic spikes.
What separates VC investing from pure gambling is the experience and knowledge of venture capitalists who use them not only to select investments but also to gently guide founders. Jervis wanted to automate that.
PitchExpert’s website states that its AI “has been trained on a variety of documents from around the web, so in many cases it can make predictions about what an investor might say that contains relevant facts about the market for your product.” .
“GPT-3 trains on a huge amount of content crawled from around the web, but it does more than ‘semantic search’ for phrases out there relevant to the topic,” Jervis told Digital Trends. “It can deal with completely new ideas and even generate useful answers to them. Deep learning models like GPT-3 are largely black boxes. Basically, statistics are used to guess what word comes next. Put simply, it turns out that if you’re writing a prompt that requires the model to guess what an expert in a given field would say, it takes a lot of work to make it convincing — and the result of that is that it’s ” Reason’ to do his job.”
PitchExpert, he explained, does not currently use any additional training data that is not part of the original GPT-3 model. In the future, however, he’s considering expanding it with additional data sources so it can ingest information about a company’s website performance, its revenue, and other metrics — and provide more tailored advice based on that.
Really valuable advice
In the case of my water bottle drone delivery system, PitchExpert’s advice was probably way too sensible for what my very stupid startup idea deserved. While some of these were relatively generic (How do you plan to scale? Think of the value proposition), other comments were surprisingly accurate. It identified some of the available markets that would be open to quadcopter bottled water and found that one of my biggest competitors would be the not exactly underfunded Amazon.
However, other, more serious founders say they really appreciated PitchExpert’s advice. “PitchExpert gives you feedback that feels like it’s coming from real VCs,” Josh Smith, creator of keyboard shortcut service Keyhero, told Digital Trends. “I’m not interested in raising money: I’m interested in developing a useful product. For me, the value of talking to an investor isn’t in their money; it is in their advice. Having raised funds before, I can tell you that the responses felt real and were really helpful.”
All of this raises a very interesting question: is Jervis’ tool a legitimate product or, well, something saucy and subversive? “I wouldn’t use the word ‘parody,'” Jervis said.
Maybe not. Still, it’s undeniably provocative to imagine that VCs, after decades of making millions of dollars from disruptive technology investments that can automate swaths of mundane jobs, might be on the chopping block themselves. Most people are familiar with “Quis custodiet ipsos custodes,” a quote from the Roman poet Juvenal, which translates to “Who watches over the watchmen?” But how about “Who automates the automators?”
Sure, PitchExpert doesn’t have the venture capital funds to invest in startups or the Rolodex to point founders to other neighboring parties. Still, it’s an exciting start. Venture capitalists, at their root, are intermediaries that connect startups that need money with funds that startups need. And if the past few years of streamlined one-click technology innovation have taught us anything, it’s that middlemen are shockingly easy to automate.
But maybe there’s a good reason why Jervis doesn’t consider this a 100% legitimate product, nor a 100% spoof. Like a kind of Schrödinger VC, PitchExpert can somehow exist in both states at the same time and serve both purposes satisfactorily. As Jean Baudrillard wrote of his favorite concept of the simulacrum: “The simulacrum is never that which conceals the truth – it is the truth that conceals that there is none.”
The reason shows how black mirror Work is because beneath it all we realize there is very little that separates much of today’s technology from some sort of hyperreality satire. Think TikTok, a time-consuming video feed so mind-reading attuned to what you want to watch that it’s virtually impossible to turn your back on it. Or Twitter, a previously 140-character messaging app that was at times used to conduct international diplomacy between powerful people who really know better. Or Google, a company that wants to organize and monetize all the information in the world to make it more useful.
These are high-profile ideas that, in another iteration of the multiverse, could easily be dreamed up by sci-fi writers to poke fun at Silicon Valley and what McKenzie Wark calls the “vectoralist class.”
Ultimately, we get the technology we deserve. And in a world of increasingly autonomous machines, perhaps it’s true that it’s the bots that know what we want better than we do – and help us decide which products and tech startups get the green light.
Even if disappointed they were lukewarm on water bottles delivered by drone.