A Great Meltdown from Within
I hope everyone has been following the newest saga in Silicon Valley. I started writing a Twitter thread this morning and realized this deserved a little more than one-line zingers and memes. So here I am, back to substack after a one-year hiatus. I am bummed about what’s happening with OpenAI for a few selfish reasons. First and foremost, as someone who loves ChatGPT and nearly replaced Google search with it, I am sad that the founder who enabled the building of this product was ousted so unceremoniously.
Looking back at the pace at which OpenAI has been shipping products this year, it’s clear they’d reached a tight working chemistry. At Bestever, putting together a team of 10 and getting everyone to the “Ship ship ship repeat” pace took us nearly 3-4 months. It’s hard to imagine how much work it was to put together a team of 700! As a team, once everyone finds their place & pace, it feels like synchronized swimming. It’s beautiful, quick & powerful. Shipping so quickly while also winning the market is an incredible feeling. Very few companies in the history of tech have been able to experience it at this scale. There is FB, Uber, and, more recently, OpenAI. Losing leadership at such a stage is probably the worst possible outcome. This is Uber circa 2017 all over again. Talented people will leave, the company's grand vision will weaken over time with a hired CEO, and eventually, the market will end up with mid-outcomes (Where are my self-driving Ubers, Dara?)
Growing a company is a mind game. Everyone has to believe they’re working to win. The best builders I’ve known in my life have all been motivated to be on winning teams. This is why non-profits have a hard time recruiting talented & ambitious people. The minute you say this is not about winning the market anymore, about 70% of the value adders will lose motivation to work, and that’s the end of the golden era of that company. This *really* sucks. As someone who was at Uber during the scandals and CEO transitions, I can tell you that bad morale is the fastest form of poison for a company.
OpenAI is unique in that it is not just a standalone company but also a new-age infrastructure provider (like GCP or AWS) that has ushered in an era of new AI startups (including @BesteverAI). So, every VC that has invested in any AI startup over the last year has also indirectly invested in OpenAI and has a vested interest in OpenAI's success.
As an early seed-stage startup, we can’t/don’t want to invest our resources in building LLMs. It is not core to our product but enables us to open up a new interaction modality that didn't exist before and dramatically enhances our user experience. Of course, we have all diversified our product to be able to easily switch out our providers, but having to do that as a startup is a completely inefficient use of our time and resources. The turmoil at OpenAI will now have second-order effects on the AI startup ecosystem. Their morale problems will no doubt percolate into their shipping speed/quality and, in return, turn into morale problems for the entire ecosystem. Every investor in the valley should be upset as if they’ve invested their own LP $$ into OpenAI.
Satya Nadella might have the reflexes of a startup CEO. Still, there is no way that the team they put together at Microsoft will be able to escape big company bureaucracy completely. Your benevolent emperor might have given you a kingdom, but it will still be governed by the existing rules. It is hard to believe that the OpenAI founders will feel as free and autonomous as they did a few days ago on their own turf.
So, where do we go from here? If the founders and investors of OpenAI envision a world where it continues to enable the AI ecosystem to build incredible products in the upcoming decade, they will get rid of this terrible board full of policy wonks and decels, reinstate the founders, and focus on the company's energy and morale back into the building. If that doesn’t happen, we’re back to the lull of the late 2010s, when nothing interesting was built.
The silver lining is that this implosion might create more LLMs and competitors in the market. But the cost is slowing down the most exciting phenomenon of this new decade before it has a chance to peak.
ncG1vNJzZmiZoKS8s8LAoKavoZ6Ze7TBwayrmpubY7CwuY6pZppll6eyosCMppylrJSkxK95xaumpmWnnsGptc0%3D