5 Things I Believe About the Future of AI

I also think a lot of very smart people are going to be very wrong about how this plays out. Including, almost certainly, me on some of it.

Founders, investors, classmates, and my own kids keep asking what I think. I run Alumni Ventures, one of the most active venture firms in the world, so I spend all day around this technology and the people building it. Here’s my current answer, in five parts.

1. Start with humility. Nobody knows.

The track record of smart people predicting technology is terrible. In 1998, Paul Krugman predicted the internet’s economic impact would be no greater than the fax machine’s. A few years later, the dot-com boom gave us the opposite mistake: investors assuming almost any company with a website would change the world. There was plenty of expertise on both sides. It didn’t prevent some very crappy ideas getting funded.

And you can be right about a bubble and still miss what matters most. That would probably be my biggest call: real breakthroughs, ridiculous valuations. Companies that look unbeatable and disappear. Companies most of us haven’t heard of yet that put a ding in the universe. Breathless hype in the short term, but 10x underestimating in the long term.

The technology is only one piece. Energy, regulation, competition, labor markets, and plain old human luddite-bias all push on each other at once. That’s a complex system, and complex systems humble forecasters. When someone tells me they know exactly how AI plays out — the singularity in three years, an 8.5% chance of an extinction event, or this is all bullshit — I tune out. Overconfidence tells me they don’t know what they are talking about.

My job is to keep learning, make considered bets, and change my mind when the evidence changes. “Strong convictions, loosely held.” I want enough of a POV to act and enough humility to adjust. Those can live together.

2. Disruption takes time. The demos leave that part out.

I was fortunate to be mentored by Clayton Christensen, so I see disruption — or the potential for it — in everything (to the annoyance of my kids). Try Jev, Muse, or Grok bots — Fable or Astra (sorry, Siri) — and you can see why people are excited. The first useful answer arrives in seconds. Changing people’s habits or how an organization works takes considerably longer.

History is emphatic on this. Electrification spread over decades, and factories had to reorganize how they worked to get the full benefit. Economist Robert Solow quipped in 1987 that “you can see the computer age everywhere but in the productivity statistics.” The technology shows up long before the gains do. This movie has been played more than Shawshank Redemption on TBS.

A hospital has patient records, liability, reimbursement, and clinicians who are already stretched. A school has teachers, parents, budgets, and a football game to win Friday night. A business has customers who expect the work to be right (Hello Blue Apron!). Signing up for a chatbot is easy; rewiring how these organizations actually work is the hard part — and that’s where much of the value lives. It moves at the speed of people, which isn’t that fast.

I see this as an entrepreneur. Getting a tool into someone’s hands is the easy part. Then you must decide which work should change, who owns it, how to check the output, and who’s responsible if something goes wrong. There’s a lot of management hiding inside the word “adoption.”

There’s also a lot of infrastructure to build. Take Groq, one of our portfolio companies. Our team was researching the opportunity in 2019 and invested in 2021, well before everyone was tracking them. Groq builds specialized chips for inference — the computing that happens when a trained model answers a question. Speed and cost matter enormously when you want to turn an impressive demo into a service people can afford to use all day. This stuff is only obvious in hindsight.

So, my bottom line: AI’s impact is slower than the hype, bigger than the skeptics. Both things are true at once: the timeline is longer than the demos suggest, and the destination is enormous.

3. When the world changes, you choose your posture. We lean in.

Here’s the part that’s personal rather than analytical. When a major technology arrives, everyone makes a choice — resist, wait, or engage. I’ve made the same choice my whole career: lean in, adopt early, learn by doing.

An AV Board member recently heard about how I was using AI for my birding hobby and said, “Mike, you’re such a nerd.”

Guilty.

Hobbies are actually a good place to start. You care about the subject, you have a reason to try things, and you learn what the tool can really do. I take the same approach at work: use the tools, check the work, figure out where they help, keep going. I don’t need a settled theory of the next 20 years to get value out of them today.

That curiosity matters in investing too. Our Oura story started with David Shapiro reading Matthew Walker’s Why We Sleep and sharing it with me. I encouraged him to contact the author, because sleep looked underinvested. He did, through an AV connection. Then he bought the ring, pursued the company, and found a path to its leadership through a former colleague. We invested in 2019. A book, a conversation, a product someone actually tried, and people willing to make introductions — curiosity becoming something concrete. I like that.

That same curiosity drives how we invest. At AV, I want us investing in two kinds of companies:

Impulse Space is a good example of the second kind. Launched by Tom Mueller, a founding employee of SpaceX, it builds vehicles that move payloads once they’re already in orbit. Better software helps, but the hardware still has to work in space. I like backing people who can do difficult, useful things in the physical world.

There will be failures in both categories. Enthusiasm doesn’t relieve us of the job of evaluating the business, the team, and the price.

4. What about jobs? Don’t bet against dynamism.

Whenever a technology this big shows up, the first instinct is to count the jobs it will kill. I resist that kind of static thinking. It tends to come from people who don’t really understand technology — or history.

Humans don’t think dynamically about systems very easily. We see the disruption; we miss the expansion. Economists have a name for one version of this: Jevons paradox. It posits that if you make a resource cheaper and more efficient, we don’t use less of it, we use a lot more.

I think intelligence works the same way. When useful thinking gets cheap, the number of things worth doing explodes. Markets expand at all kinds of scales and levels. AI for birding? Who’d a thunk.

Some segments will change. You don’t want to be the toll-booth-operator equivalent. But I believe much, much more gets created than destroyed. And look at what’s actually happening. AI has been in the economy for a while now, and I’m seeing more — more companies, products, and work worth doing — not less.

Here’s my call: jobs get broader and more interesting. Liberal arts are back. People who think like entrepreneurs and CEOs are superpowered. People who treat a job as a static set of workflows that never touch another human — their world is going to change, and not for the better.

The upshot is that this technology actually lets us be more human. No kid woke up in fifth grade and said, “I want to sit in a cubicle and type into a keyboard all day.” Good riddance. What’s left is the stuff we’re actually built for:

That’s where the magic is, letting humans be more human. Sign me up.

5. The reason for optimism: abundance.

The strongest case for optimism isn’t what AI does for people like me. It’s what it can do for the parts of the world that still aren’t good enough: education, healthcare, poverty, opportunity, freedom, agency.

Think about a kid who needs help with math and whose family can’t afford a tutor. Benjamin Bloom’s classic research showed how much better students do with individual attention. The problem has always been making that affordable for everyone. AI may finally put a patient tutor in every kid’s pocket, in every language, for almost nothing. Khan Academy effectively does it already.

We still must prove what works, and there’s the diffusion issue. While that’s happening, people overcorrect. But the payoffs are already appearing. In science, AlphaFold has predicted structures for some 200 million proteins — work that earned Demis Hassabis and John Jumper a share of the 2024 Nobel Prize in Chemistry. But we still have to change that prognosis for the patient who was just diagnosed with Stage 4 Cancer.

Closer to home, I have a text chain with friends and family called “100 and healthy,” where we share what we’re learning about longevity. The “healthy” part matters. That’s one reason portfolio companies like Oura and Function Health interest me. Both are early examples of AI turning personal health data into something you can act on. Oura puts sleep and other health information into people’s daily lives. Function combines diagnostic testing with tools that help people understand their results over time.

More AI-powered information, earlier, with a chance to actually do something about it. I want more healthy years with the people I care about. So does everyone I know.

I don’t assume the benefits distribute themselves evenly (a visit to a family reunion will tell you that). And I take the headline fears seriously. Misuse by bad actors, systems we don’t fully control, the energy appetite of all this compute. History says big technologies arrive ungoverned and we build the guardrails as we go: electricity, cars, aviation, and the internet all followed that arc, imperfectly but genuinely.

Some of the best answers will come from builders. At AV we look every day at companies whose work bears on these problems, from security to the energy buildout. Those are real responsibilities that belong in the conversation alongside the possibilities.

But over generations, better tools and scientific progress have helped people live longer, learn more, and escape poverty. Expertise has always been expensive and scarce. AI could make much more of it available for much less — and abundance matters most to the people who had the least access to begin with.

Where I come out

My mindset: Humility about the path, patience about the timeline, and a decision to lean in. And real optimism about what abundance can do for the problems that deserve it most.

I respect people who come at this differently. Parents, regulators, teachers, founders, and investors carry different responsibilities and should ask different questions. My answer is also written in pencil, like any honest view of something this big.

For me, the next step is straightforward. Keep using the tools, making AV AI-first and AI-proof, backing builders doing the same, and paying attention to what’s actually working. And staying curious, learning, leaning in.

If you’re inclined to lean in too, staying close to AV and our companies is a hell of a front seat.

Further viewing: Pat Grady on the next wave of AI

For another perspective from the venture community, Sequoia Capital investor Pat Grady looks at the pace of AI progress, the gap between new capabilities and real-world adoption, and the opportunities and risks he sees for startups.

This communication is from Alumni Ventures, a for-profit venture capital company that is not affiliated with or endorsed by any school. It is not personalized advice, and AV only provides advice to its client funds. This communication is neither an offer to sell, nor a solicitation of an offer to purchase, any security. Such offers are made only pursuant to the formal offering documents for the fund(s) concerned, and describe significant risks and other material information that should be carefully considered before investing. For additional information, please see here. Achievement of investment objectives, including any amount of investment return, cannot be guaranteed. Co-investors are shown for illustrative purposes only, do not reflect all organizations with which AV co-invests, and do not necessarily indicate future co-investors. Example portfolio companies shown are not available to future investors, except potentially in the case of follow-on investments. Venture capital investing involves substantial risk, including risk of loss of all capital invested. Diversification cannot prevent investment loss; it is a strategy to mitigate investment risk. This communication includes forward-looking statements, generally consisting of any statement pertaining to any issue other than historical fact, including without limitation predictions, financial projections, the anticipated results of the execution of any plan or strategy, the expectation or belief of the speaker, or other events or circumstances to exist in the future. Forward-looking statements are not representations of actual fact, depend on certain assumptions that may not be realized, and are not guaranteed to occur. Any forward-looking statements included in this communication speak only as of the date of the communication. AV and its affiliates disclaim any obligation to update, amend, or alter such forward-looking statements, whether due to subsequent events, new information, or otherwise.