Larry Kim, Founder of WordStream: Which Skills Will Still Matter in Five Years After AI Takes Over Everything
Larry Kim, founder of WordStream and CEO of Customers.ai, poses the question: What skills will still be important five years from now after AI takes over everything? As a father of two young sons, he shares his plan—encouraging his younger son to pursue relationship-based sales and guiding his older son towards becoming an electrical or mechanical engineer, with a goal aimed at SpaceX.
Kim founded the search advertising automation company WordStream in his twenties, which was acquired by Gannett for about $150 million in 2018, at that time employing around 300 people, generating over $55 million in annual revenue, and serving over 10,000 global clients. He claims to have raised only about $18 million in funding. He later started a new venture focused on website identity recognition and remarketing, currently residing in Cambridge and Cape Cod, Massachusetts, with a background in electrical engineering from the University of Waterloo, and became a U.S. citizen in 2010.
He has publicly introduced his two sons: one about 6 years old who likes insects and lizards, and the other about 10 years old. An interview mentioned that his older son struggles more on the autism spectrum, and he hopes his child will eventually enter software entrepreneurship. Recently, he also posted about his 8-year-old child using generative tools to write large database applications, transforming "knowing how to code" from a professional skill into a task that can be handled by models.
He categorizes the skills that will still be relevant in five years into two types: one that cannot be completed solely by models, which involves face-to-face transactions and long-term relationship maintenance; and the other that still requires assembling rockets, satellites, and electrical systems in the physical world. SpaceX is specifically named as a concrete employer, rather than the abstract term "STEM."
In market mechanisms, what is bought is an educational option for "skills still scarce in the physical world," while what is sold is a hedge against pure knowledge work being replaced by models. Funding and time are shifting from general programming training to sales practice and hard engineering disciplines. Benefiting are positions that can still establish trust, on-site delivery, and aerospace-grade hardware as thresholds; under pressure are families treating writing, placement, analysis, and basic programming as ultimate skills. This is a career hedge made by parents using two samples, not an industry recruitment announcement.
Source: Public Information
ABAB AI Insight
Kim's first pot of gold came from automating repetitive tasks in advertising placement and selling it to media groups. Now he questions what skills will remain in five years, using his own automated industry as a cautionary tale: WordStream eliminated the need for optimization specialists, while Customers.ai replaced the human layer in visitor identification. He is not pulling his sons out of "making another version of marketing software" by chance.
The capital path is to use cash from an exit to buy career insurance for the next generation. The $150 million acquisition provides a trial-and-error budget for the family, allowing the younger son to be directed towards high-contact sales and the older son towards aerospace engineering, which requires clearances, workshops, and security checks. The motivation is to hedge against the field he knows best—digital marketing and software—being thoroughly consumed by models. The strategy is to place the children in two types of interfaces that AI cannot replace in the short term: trust between humans and machines, and assembly between bits and atoms.
Analogous to Silicon Valley parents sending their children into careers in poker, real estate development, or pipeline businesses, and considering "going to SpaceX" as the new "going to NASA." The industry phase is one of accelerated replacement: the entry barriers for software positions have been flattened by generative tools, while the barriers for hardware and face-to-face transactions remain temporarily intact. Kim includes SpaceX in his plans because that company requires electrical, mechanical, and highly collaborative on-site work, rather than relying on the rockets themselves to guarantee employment.
Structural judgment belongs to technological replacement. What is being replaced are jobs that can be fully described by text and code; what remains are jobs that must be present in the room or on the launch pad. The mechanism is that models first consume standardized outputs, then judgments that can be delivered remotely, and only later do they encounter positions that must bear personal and physical consequences. Relationship sales and aerospace engineering are temporarily stuck at the third level, thus chosen as family hedges.
ABAB News · Cognitive Laws
- Models first consume skills that can be delivered remotely, then those that must be present.
- Parents' career planning is a vote on the speed of replacement in their own industry.
- Relationships and atoms are the two gates left when bits can no longer be consumed.