How generative AI is changing startups

With the next generation of AI tools, teams of three very talented people will be able to grow software-centric businesses to $100+ million in revenue with automated workflows.

They will be able to set up AI systems for each aspect of their business and let them run, like spinning plates, periodically revisiting them to improve them and keep them spinning.

They will be able to develop software faster and better with AI dev co-pilots. Run sales prospecting, qualifying, and outreach with AI automated systems. Run marketing campaigns with AI optimizations. Run AI customer service and success faster and with higher quality. Run accounting and legal cheaper and faster. Run analytics with more detail, less fuss, and better results. Set up self-healing data pipelines. Set up automated workflows. File taxes and other government requirements. All with AI.

Founders in the era of AI will be able to do a lot more with a lot less.

Allometric Scaling and Generative AI

The reason is allometric scaling, and it compounds.

In biology, allometric scaling is the idea that there is a relationship between certain characteristics of an organism (organ size, heart rate, metabolism, lifespan) and size. Some things go up relative to size, other things go down.

Allometric scaling is at play in every company in the generative AI era. You’ll notice lots of different ways that AI alters the relationship between your company’s size and its metabolism, growth, heart rate.

Some things won’t change. Smaller teams will always move faster, relative to larger teams. But other characteristics will be very different. The formula we use to calculate revenue per employee will change, for example.

If you can have fewer people, every business operation gets easier, and as business elements get easier, you need even fewer people. Fewer people means less of everything that isn’t building and selling to customers.

Fewer people means founders have fewer meetings, fewer disagreements, less politics, faster decision making, and faster experimentation.

Fewer people means less recruiting, interviewing, hiring, onboarding, coaching, performance reviewing, culture building, dramatic departures and firings.

Fewer people means less office space, office rules, or squabbles.

Fewer people means less salary spend, less fundraising time spent, less capital raised from VCs and less dilution to the founders.

This isn’t far-fetched. If you recall, we’ve already seen this alternative allometric scaling in tech companies before AI. It was just isolated in a small, but very visible, niche of the tech world: social media. Instagram had 12 employees when it sold for $1B, and WhatsApp had 50 people when it sold for $21B.

With AI, the TYPES of businesses that can achieve this type of allometric scaling will increase dramatically.

To be fair, to pull off a three person unicorn will require very special talent among the three people. Few teams will have the multifaceted skills to pull it off.

When is this happening? Now.

Early Evidence This is Happening

We’ve already seen tools that allow AI to write its own to-do lists, assign tasks, write a hundred different software versions of each new task, and test each one. AI is creating and testing new simple features, largely without a human in the loop already.

We’ve seen customer service organizations drop 90% of their people after implementing good AI-enabled software.

We’ve seen both software engineers’ and digital artists’ (in games) output increase 100% with the right co-pilot AI tools.

Generative AI has become consensus inside companies big and small, as we predicted. Already 93% of NFX founders are using ChatGPT on a daily basis, and 95% believe generative AI is either a valuable accelerant or a gamechanger for their specific industry.

In concert, we’ve seen decrying about potential job loss begin. The common phrase right now is “AI won’t replace you, but a human with AI will.” That’s already happening.

Recruiting is about to be 90% automated. You will be able to describe your ideal candidate to an AI scribe which will write the job descriptions, distribute the posts, rank the highest quality candidates, and book time on your calendar to interview them. The AI scribe would record the interview, and then make a hiring recommendation based on all data collected. Recruiting will look entirely different.

[ Note: It’s the same for VCs too. So much of what a typical venture capitalist does – reading writing, summarizing, and ranking – is also what LLMs do extremely well. The more data on hand, the more AI will eclipse the humans in the loop. Growth stage VC will be affected first. Seed, where intangible human elements are more present, will be affected last. ]

AI is going to beat most people’s average performance

Generative AI is going to outperform most humans, on most hard skill tasks, on most days. Not all days, but most. AI is just going to be good more consistently. The result is that, for the information work that we’re concerned with, the defensibility of hard skills declines over time.

On a day to day basis, even those of us who are capable of great, aren’t being great. No one has an outlier day, every day. Most of what Hemingway wrote on a day to day basis ended up in the trash.

AI, though, is capable of regularly maintaining a very high average performance, if not a better average than humans at specific, hard skills. AI eliminates the random fluctuations of organic life. No more bad days.

It’s More About The Outliers Than Ever

What does this mean for you? Founders of the 3-person unicorns will need to be prone to a stroke of genius, and unafraid of acting on these impulses. You need to be capable of seeing something others don’t, and aggressive, brave or disagreeable enough to take action.

You have to be able to get out of your own way.

For the next few years at least, AI is likely to fall short in the realm of the extraordinary statistical anomalies and outliers.

What does The 3-Person Unicorn look like?

AI is facilitating a move from individual and specific expertise, to generalist skills and the ability to product-manage AI tools. All will need to have project management skills, but each might have a talent and personality slant:

1. The Founder: Aggressive Visionary Archetype
2. The Numbers Person
3. The Words Person

Everyone must be capable of juggling multiple roles. All have to be capable of outlier performances.

1. The CEO: Aggressive Visionary Generalists

Because of the shrinking moat around hard skills, the archetype for successful founders in the next wave of generative AI-run startups is: the aggressive, visionary generalist.

The “aggressive visionary” part covers the personality and mental model you need to succeed as a Founder. The “generalist” part is the core skill that allows you to flexibly work with AI.

Founders need to know if they’re in fast moving water, that their idea is in fertile ground, and have an idea of what it takes to be defensible in your field, in the face of generative tech.

Emily LeProust, the founder of Twist Biosciences and part of the NFXBio team, is an elite example. Emily wrote the patents that made Twist Biosciences. She raised the money. She did the initial sales. She built the product. She led the hiring. She’s savage in the best way.

2. The Words Person

The humanities just got an upgrade in status.

Your words person can lift you to the highest levels in your sales and marketing. They have the empathy and nuance to pick the right language from everything the AI models generate. They understand psychology. They no longer need to be able to write great stuff from scratch anymore, they just need to know it when they see it.

3. The Numbers Person

These folks think in numbers and metrics. They love spending a Sunday morning pouring through database queries to find patterns and anomalies. They are uncomfortable when something is described without a number attached. They are precise.

AI has shifted the advantage to the small cell

AI gives more power to a small group than ever before. If you have the world class team to pull off a 3-person unicorn, we believe in you.