AI Q&A: Five Questions for the Road Ahead
Kevin Laws '88 talks education, careers, social order in the AI age
Kevin Laws ’88 has worked at the leading edge of technology since the release of the world’s first web browser. His expertise at the intersection of technology, investment and business informed his success as CEO of AngelList—the revolutionary platform matching startups with investors—during the company’s major expansion. It also drives his current work investing in and advising AI companies, as well as his upcoming position as executive-in-residence at Oxford University, beginning in fall 2026.
The text below has been lightly edited.
1. What are some ways AI can benefit society?
The benefits are substantial: AI reduces the friction between having an idea and making it real. This works most powerfully where reality is digital—software, web services, gaming. But it extends further than you might think—into design, research, law, accounting and many more forms of knowledge work. In my field of turning ideas into products that serve others, AI has been transformative. Teams can now launch products with literally just one or two people.
2. What are your concerns about AI?
I don’t worry that AI will become sentient or put everyone out of work. I worry because many people misunderstand how it works and often misuse it, because they fail to recognize its limitations and its tendency to reinforce existing biases.
Beyond that, the challenges are more subtle than the "evil-genius AI" scenarios many people focus on. We've built our entire society around the idea that certain skills are hard and rare and that other skills are not. AI disrupts those assumptions.
We've also over-rewarded certain skills. Computer programmers fresh out of college can make literally four times what people in some other professions earn. Adjusting to those skills becoming less valuable—while skills like connecting with people or working with your hands become more highly prized—will shock our society. I hope all these shocks to so many parts of our society happening at once don't tear us apart.
3. You developed an AI tool to help high-schoolers with college searches, and you've taught at the college level. Tell us your thoughts about education in the AI age.
For the first time, we have technology that could genuinely accelerate parts of education—transmitting knowledge, frameworks, heuristics. This forces us to reconsider what aspects of education are truly valuable. I suspect it will turn out that mentoring, socialization and presenting a stream of increasingly more difficult life challenges will be more important than we thought. I also suspect that AI may make that harder, not easier, as students use it to avoid the hard challenges.
AI will make skilled professionals incredibly productive—lawyers, accountants, programmers, consultants. But here's the problem: The next generation of those professionals traditionally learned through apprenticeship, by doing the "boring work anybody could do" while observing experienced professionals. If anybody can do that work, so can AI. How, then, does the next generation learn?
4. What's your best advice to young people who want to work in the tech sector?
I’d ask them to think about what they are passionate about and how they might turn their passions into something valuable to others. I’d also ask them about whether they’d best serve that cause as creator or as one who helps others create it. Mistakes are inevitable—and ultimately helpful.
I’d also advise them to avoid focusing solely on grades. The most interesting people I work with rarely have perfect transcripts. They have interesting failures that they were clever and persistent enough to turn into success.
5. What’s the most important thing you’ve learned about AI?
People simultaneously overestimate and underestimate AI. They overestimate it when they imagine current LLMs will become sentient and take over the world. These systems are essentially sophisticated autocomplete and translation tools. They underestimate it when they miss how powerful translation and autocomplete can be. After all, it's "just translation" to turn English to computer code, and you can autocomplete things like "If I wanted to make genetic changes to a basil plant so it doesn't turn brown, I should … "
Learn more about Kevin Laws '88 and his path from Silicon Valley to Oxford.
Read more about Dickinson alumni working in AI.
TAKE THE NEXT STEPS
Published January 26, 2026