“The development of AI is as fundamental as the creation of the microprocessor, the personal computer, the Internet, and the mobile phone. It will change the way people work, learn, travel, get health care, and communicate with each other. Entire industries will reorient around it. Businesses will distinguish themselves by how well they use it.”
–Bill Gates, GatesNotes, The Age of AI has begun, March 22, 2023
We believe that this new generation of AI is the most important development in technology since the Internet. It will impact every industry and every job function. As Bill Gates writes, “It will change the way people work, learn, travel, get health care, and communicate with each other.”
This is not crypto. The changes that AI will bring (and is already bringing) and the opportunities it creates for new technology companies are profound and real.
This AI change will take on two main forms: augmentation and replacement. That is, AI will augment human functions and, yes, replace many of them. In the immediate term, it is unclear the extent to which machines will replace certain job types and functions and the extent to which machines will simply augment those jobs and functions.
In the longer term, we believe that inevitably the jobs and functions that machines do better will be done by machines, the jobs and functions humans do better will be done by humans, and the jobs and functions humans do better when augmented by machines will be done by humans assisted by machines. This new type of AI will ultimately affect every job and every function in a meaningful way.
The social and political implications of AI are vast. What will happen to the people who currently occupy the jobs that machines replace? Will this AI lead to more wealth disparity and political stratification? Can we as a global society put the right regulations and technology parameters around AI so that it serves the interests of humankind versus the interests of bad actors or itself?
These are important issues, ones that we hope to play a small role in addressing through rapid company-creation and working with our legislators on policy-making. In the meantime, we have jumped into the deep end of the AI ocean – no paddle, no floaties, just trying to learn as much as we can as we work our way through the middle of the muck.
What are we doing with generative AI now?
We have been working on AI-driven innovation since the studio launched eight years ago. In fact, the vast majority of PSL’s portfolio leverages AI or machine learning (ML) in some important way. In mid-2020, when GPT-3 first launched, we built some prototypes and internal tools to take advantage of the new generative capabilities.
The game changed when ChatGPT launched in November 2022. One of our principal engineers, Kevin Leneway, explained in a blog post he wrote on PSL’s site in June 2022 that using the 2020 version of GPT-3 felt like “giving a 12-year-old a Ferrari” and that “it looked really cool, and you just knew that in the future it was going to be awesome, but there just wasn’t much that could be done with it today.” Fast forward three years and ChatGPT handed all of us a learner’s permit to drive that Ferrari. For the first time, we’re able to leverage the power of these LLMs to build generative applications that perform human functions in response to natural language inputs in a way that wasn’t possible only months ago.
We are currently building multiple apps at PSL using this new generative AI capability. For example, we have constructed a co-pilot for product managers in the hardware space, a compliance co-pilot for finance professionals, a co-pilot for healthcare administrators, and an agent for coding more complex production applications. Even in the midst of building all these applications, we are still trying to figure out ways to go even faster and learn even more about this new capability..
Our entire studio team uses generative AI extensively and daily. For example, our engineers are using GitHub Co-Pilot and GPT-4 to write code and claiming a 2x to 3x efficiency improvement. This post is human-generated, but we will run it through ChatGPT to edit it.
Every single one of our portfolio companies is figuring out how to take advantage of generative AI to make their employees more efficient and, in some cases, to rethink their product offerings to customers.
In the past several months, our venture fund has made three new investments in non-studio companies with world-class teams that are leveraging generative AI in the security, defense, and education verticals.
The cost to create a company that leverages this type of AI is substantially less than it was to create a company with similar capabilities a year ago. A handful of developers can build magical experiences on top of the new LLMs, and there are new companies being formed now to address almost every industry vertical and function. The large technology companies, Microsoft, Google, Amazon, and Meta, are devoting billions of dollars to building fundamental infrastructure, making it hard to compete in the underlying technology.
So how will we differentiate?
To attract great founders in this space, our studio needs to be the best in the world at rapidly building new AI companies at the highest quality possible. We will need to leverage the many advantages we already have and be creative and aggressive in trying new company-building approaches. In the coming months, you will see us testing new models of company-building with the goal of building more companies more quickly and learning more from the many great entrepreneurs in this space.
We are fortunate to be based in Seattle where some of the very best AI engineering talent resides. Microsoft, Google, Amazon and others have large AI centers of excellence in the Seattle area and arguably the best talent in the world in this space. The University of Washington, where several of us serve as adjunct faculty members, is one of the top universities in the world for developing AI talent. AI2 is a leading global research institution and incubator in the space. Our goal is to partner actively with all of them in win-win ways to create market-leading companies.
As mentioned earlier, we also are partnered with and building startups directly alongside some of the world’s leading companies. They have unique, proprietary data sets, which lead to sustainable differentiation when combined with an LLM. These companies also are interested in building generative AI applications in their respective verticals with talented outside entrepreneurs. Stay tuned for more exciting developments on this front.
Product Experience and Business Model Still Matter a Lot
When transformational change like generative AI takes center stage, a knee-jerk reaction for some entrepreneurs is to go broad with a solution that solves everything (and nothing very well) and believe that the underlying technology will make up for product deficiencies.
We take the opposite approach. We start narrow, trying to be the best in the world at something small and tractable. We want to make it irresponsible for early customers to choose another solution to solve their particular, circumscribed problem.
We lean heavily on our best-in-class design team to build UX experiences that are functional but also delightful. In a crowded market, where the foundational technology is built by others, better UX and visual design are two areas where PSL can help entrepreneurs differentiate their solutions.
Having invested in many hundreds of companies as a team, we also can help founders think strategically about their business models and the milestones necessary to achieve funding and product-market fit.
Even though small technically-minded teams can do more now than ever leveraging generative AI, recruiting A+ talent is still the single most important factor for successful startups. We have placed hundreds of candidates in positions at our startups across levels and functions. We are confident in saying no one can help more when it comes to recruiting, especially in our backyard. Helping companies recruit great talent will always be our secret sauce.
The Road Ahead
With equal doses of humility and determination, we aim to build a studio that is worthy of our founders, investors, and the once-in-a-generation opportunity that stands before us.
We are grateful for and excited about the road ahead. We want to leave you with one question: Looking ahead 20 years, is it more likely that the top 50 technology companies will remain in place or that this AI revolution will usher in a new set of technology leaders that replace many of today’s incumbents?
We believe the latter and, in the words of Walt Whitman, that as “the powerful play goes on” during this transformational time in history, we might “contribute a verse.”