The Twenty Minute VCElias Torres, Co-Founder and CEO @ Agency: What No One Tells You About Selling Your Company
CHAPTERS
Drift as a “$1.2B failure”: ego, relevance, and what founders really want
Elias opens with a raw take: despite a massive outcome, Drift still feels like the biggest failure of his life. He frames success as building something enduring—where the company can thrive without the founder needing to matter.
- •Selling can leave founders feeling irrelevant inside the company they created
- •Enduring/sustainable companies matter more than flashy outcomes
- •Founder ego is a real cost of management changes and exits
- •Regret framed around giving too many “carrots” vs building for durability
“I don’t care” vs caring too much: coping, resilience, and money’s effect on stress
Harry probes Elias’s tweet about pretending to be a nihilist. Elias explains the paradox: he cares deeply, but experience (and financial security) reduces panic about many day-to-day problems.
- •Entrepreneurship toughens you through rejection and churn
- •We often overestimate how hard things are; starting is easier than finishing/exiting
- •Money doesn’t remove caring, but it removes many material stressors
- •Core mindset: things will be okay, and obstacles are often self-imposed
From Nicaragua to the US: homelessness, the American dream, and immigrant grit
Elias recounts the revolution in Nicaragua, losing housing, and arriving in the US via a green card enabled by his grandmother. He describes the US as still uniquely opportunity-rich and ties immigrant backgrounds to startup grit.
- •Revolution and property loss led to literal homelessness after high school
- •Green card timing forced a decisive move to the US
- •Deep optimism about opportunity in the US remains intact
- •Immigrant adversity can create higher tolerance for discomfort and risk
First exposure to computers and IBM: early internet building—and why big companies can’t move fast
Elias describes discovering computers via an old IBM PC and then working at IBM on early web applications. He argues that scale and bureaucracy make large companies fundamentally slower, despite claims that “elephants can dance.”
- •Early fascination: WordPerfect/Lotus 1-2-3 on an IBM machine
- •IBM internet-era work: early HTML/JS browser-based apps (1998–99)
- •Core claim: big-company headcount and process create inertia
- •Disagreement with the idea that large orgs can truly “dance” like startups
AI incumbents: distribution isn’t speed—and why “operatorless” software is the real shift
Harry challenges Elias on whether Google/Microsoft/Adobe are moving faster than ever. Elias argues it’s mostly talk, and that the real future is software designed for autonomy—like driverless cars—rather than tools requiring human operators (like ChatGPT today).
- •Incumbents’ advantage is distribution/data, not agility
- •ChatGPT is a major step, but still makes the user the operator
- •Future software should be built for agentic execution, not UI-driven workflows
- •Big bet: autonomous workflows change how B2B tools should be designed
Why management breaks at scale: humans as the blocker, small teams, and CEO-led adoption
Elias explains why organizations slow down: approvals, incentives to say no, promotion politics, and layers of management. He predicts smaller teams can reach massive revenue by substituting labor with compute—if CEOs personally drive adoption rather than delegating it down the org chart.
- •“Nobody gets fired for saying no” reinforces stagnation
- •One-on-ones and meeting-heavy cultures can kill execution speed
- •AI shifts cost structure from labor to compute; enables smaller teams
- •Enterprise adoption requires CEO ownership; middle layers resist change
Drift’s post-mortem: why a high-profile outcome still feels like failure
Elias calls Drift his biggest personal failure and pinpoints the cause: too many people, slowed product execution, and missing the timing of the LLM wave. He contrasts “enduring product value” with scaling headcount and losing speed.
- •Failure defined by durability and customer value, not valuation
- •Headcount growth created inefficiency and loss of product focus
- •Speed/timing mattered—LLMs arrived just after the window he needed
- •Lesson: avoid letting customers down by losing execution discipline
Selling the company: ego, realism, and what exits don’t fix
The conversation turns to the sale process and life after an exit. Elias describes the emotional hangover—becoming “nobody,” losing relevance—and how wealth changes logistics but not identity, purpose, or relationships.
- •Sale decision framed as pragmatic, not emotional
- •Peer valuations can distort ego and decision-making
- •Exit aftermath can include depression, loss of meetings/status, identity drift
- •Wealth enables focus (assistants, fewer errands) but doesn’t solve meaning
How to hire “rockstars”: high-agency traits, intensity as a filter, and proving by doing
Harry shifts to Elias’s reputation for recruiting elite talent. Elias lays out his criteria—high agency, IQ, grit, speed, obsession—and describes a high-pressure interview style (e.g., show GitHub now) designed to filter for performers who thrive under intensity.
- •Hiring criteria: high agency, execution speed, grit, obsession
- •Interviewing is a stress test—intensity is intentional
- •Preference for engineers who can talk to customers and act fast
- •Use contracting/part-time trials to validate performance quickly
Hiring mistakes founders repeat: desperation, slow firing, and the bounce problem
Elias explains that many mis-hires are self-inflicted—founders see warning signs but rationalize them due to urgency. He advocates hiring slower, firing faster, doing deeper references, and not tolerating job-hopping without clear slope and substance.
- •Mis-hire root cause: founder desperation and self-deception
- •Hire slow, fire fast; most founders delay firings due to ego
- •Trial work reduces risk and reveals true throughput
- •Beware “bouncers” and story-heavy candidates; probe for specifics and data
Titles, incentives, and the new company model: profitability, legacy, and VC expectations
Elias argues titles are mostly ego artifacts and that companies need new reward systems (profit/revenue sharing, equity, compensation redesign). He also discusses the shift from ZIRP-era growth-at-all-costs toward durability, and how AI may change venture and financing structures (compute vs headcount).
- •Titles distort behavior; seek people less motivated by hierarchy
- •Re-think rewards: profit sharing, revenue sharing, equity, higher cash comp
- •AI-driven companies may need less equity-heavy growth funding; more debt later
- •Balancing profitability and growth is a paradox—often “both,” not either/or
Everything you think you know about working parents is wrong: kids early, marriage, and trade-offs
Elias defends having children earlier, arguing biology and societal trends make waiting risky and that it’s possible (though ugly) to build and parent simultaneously. He shares lessons on marriage, parenting styles, love languages, and the unavoidable feeling of letting people down.
- •Case for kids earlier: biology, IVF difficulty, societal demographic decline
- •Founder performance vs family demands—redefining what “peak” means
- •Marriage advice: reduce selfishness; social media misleads expectations
- •Parenting paradox: be loving/affirming while manufacturing adversity
Operator vs investor: why Elias chose to build Agency (and how consulting shaped it)
Elias explains why he didn’t become a VC: he’s a builder who dislikes advising from the sidelines. He describes earning significant income consulting on LLM implementations for major brands, which helped crystallize the idea for Agency and his view that services will matter during the adoption gap.
- •Investing is hard; operating is more fun and aligned with his identity
- •Consulting for OpenAI customers revealed real enterprise adoption blockers
- •AI services may boom in the transition period before true agents/workflows
- •Adoption will likely take longer than hype suggests; we ‘just discovered fire’
Quickfire beliefs and closing optimism: directness, extremism, and building better craftsmanship
In rapid-fire, Elias shares contrarian views (directness, NVIDIA as a compute bet, shorting Salesforce) and personal reflections on happiness and parenting. He closes with optimism inspired by Egyptian craftsmanship—hoping AI enables better products, better customer relationships, more art, and more family time.
- •Belief: direct action and bluntness beat indirectness
- •Concern: extremism and inability to have real conversations
- •Compute demand as a long-term theme; skepticism of many incumbents
- •Optimism: return to craftsmanship—better products, culture, and time with family
