Each year, when the summer ends, I begin to think about what I’ll write in this annual “breakout company” post, which has turned into a mini-annual tradition for me here on this blog. Until a few days ago, I kept thinking – just like in 2015-16, I don’t think I’ll be able to declaratively write, “this company is breaking out.” Over the last 15-year bull run, entrepreneurs and VCs all but harvested the consumer spending wallet share, with travel/transport, food/delivery, and nearly everything else imaginable. As a result, the technology sector has, for the past few years, been eagerly awaiting the arrival of the next great platform only to be somewhat disappointed.
That changed this week with a new product release. Each year at this time (well, all year), I text voraciously with my old friend Josh, where we debate this post (and he should get some extra credit here for that). For this year’s pick, Josh summed it up nicely in text: This year’s pick “gives us the most exciting glimpse into the future.”
So with no further ado, The Breakout Tech Company of 2022 is: OpenAI.
Now, is OpenAI technically a tech startup? Isn’t it a foundation, or a capped profit company? Didn’t they raise hundreds of millions of dollars? Frankly, I don’t know how to classify it. It is not a conventional pick (before you send me a text about that), but 2022 wasn’t a conventional year, either. Before we get into it, I wanted to note that honorable mentions are warranted this year, companies such as BeReal (a super fun product), Handshake (quietly becoming next the LinkedIn), Replit (a fascinating vision for the future), Hugging Face or Stability.ai (insane early growth), for instance, were on my mind as candidates, but the spirit of this post is to isolate the newest company that is putting technology into peoples’ hands at the fastest rate – this year, I couldn’t find anything that held up to OpenAI. I don’t want to get into technicalities here this year on this dimension, given there wasn’t a clear, traditional choice, and also the overwhelmingly positive and potentially-game-changing implications to OpenAI’s signature releases this year.
OpenAI was originally formed a few years ago with the intention of investing in engineering and design talent, learning models, compute, and anything else that could help accelerate and usher in a new era of responsible use of artificial intelligence. We won’t get into the debate on the merits or dangers of AI, as that’s for another post — though I’m wholly unqualified to opine on things like TayTweets. (The leader of OpenAI, Sam Altman, needs 1-2 sentences of credit here – after hyper-scaling Y Combinator in a controversial yet, ultimately, directionally-correct move, he jumped to OpenAI in 2019 and has helped lead the organization to this point. That’s an impressive resume.)
Since then, OpenAI has been shipping. Major releases by OpenAI include GPT-3 in June 2020, a machine learning toolset that demonstrated the first glimpse of what learning models could do to generate text. In 2021, OpenAI released Codex, an AI platform that turns natural language into code, a huge advancement of making software development faster and more accessible. Fast-forward to this past fall, OpenAI released Dall-E 2 to incredible fanfare, a deep learning model that generates digital images from natural language descriptions. Dall-E 2 imagines took over social media for days, along with other tools such as Midjourney. During these releases, OpenAI led an investment in Descript, teasing a possible future of audio generated by AI, and ultimately all sorts of media. Then earlier this week, OpenAI released ChatGPT, a dialogue-based AI chat interface for its GPT-3 family of large language models. This release again took over social media in tech circles and harkened back to the craze around bots 5-7 years ago – and frankly prompted Josh and I to finally text me back with “yeah, OpenAI is the breakout.”
The spirit of this tradition is not to identify what will be the next great company. The fact is, we don’t know what happens moving forward. It’s not clear if closed-sourced or open-sourced models win the day. It’s not clear if audiences will ultimately want machine-generated communications, or if regulators will step in. It’s not clear if moats will exist as the compute costs of building and managing these models will stay high or drop. It’s not yet clear if AI-native companies can disrupt and leapfront tech incumbents, as Elad Gil artfully argues; but, it’s also not year clear if incumbents, such as Notion, can more easily “layer in” AI-driven user experiences into its established platform.
In this post, I am trying to isolate one company that placed the most new technologies and/or products in peoples’ hands. The tech/startup sector has been dreaming about the arrival of AI for decades, with the chorus growing louder over the past few years. Like many technological advances, it feels like it’s happening very slowly, and then all of a sudden, it’s here. That’s what OpenAI and others have done to the field of artificial intelligence in 2022. Most players in the startup ecosystem have been eagerly awaiting the arrival of the next great platform, only to be left waiting or running out of funds. In contrast, OpenAI has demonstrated the ability to show what these kind of platforms can do. The use cases are infinite, while the trajectory forward and business models remain wholly uncertain.
It wouldn’t be 2022 without a sober note of warning. It’s now been 10 years that this blog has picked a “breakout.” Let’s look over the past ones – OpenSea from 2021 is today facing major crypto headwinds; Hopin from 2020 is today struggling to stay relevant as the public now mostly rejects online events post-pandemic; Superhuman from 2019 is today likely flat with respect to growth; Airtable from 2018 is today still growing but perhaps will face a big valuation adjustment given public comps; Coinbase from 2017 is today taking a beating in public markets given the icy crypto winter we’re in; there were none in 2015 & 2016 (at the time, in my opinion); Slack from 2014 has since been subsumed by Salesforce though I’d speculate is still growing modestly; Snap from 2013 is today battered in terms of its market cap despite having strong user engagement and revenues; and Stripe from 2012 remains the most dynamic private tech company at scale, but missed its IPO window given the downturn, and could be valued closer to Snowflake upon going public – which is still huge.
The meta point here isn’t to talk about the market troubles here, but rather to point out that while the startups selected annually for this post have broken out, breaking out in and of itself does not ensure longevity and endurance. Technology is moving faster than ever, and while OpenAI is moving really, really fast right now, something else that’s fledgling today with promise may break out to disrupt one or more of these companies next year.