The Twenty Minute VCElias Torres, Co-Founder and CEO @ Agency: What No One Tells You About Selling Your Company
CHAPTERS
- 0:00 – 5:32
Caring too much, pretending not to: ego, stress, and what money actually fixes
Elias opens with a paradox: he projects nihilism to cope with how deeply he cares, especially as an entrepreneur constantly facing rejection. He explains how wealth reduces day-to-day stressors but doesn’t remove the emotional weight of building and being judged.
- •Entrepreneurship trains emotional resilience through constant “no’s” and churn
- •The ‘I don’t care’ posture is a coping mechanism, not true apathy
- •Money removes many practical anxieties (repairs, fees, logistics) but not existential ones
- •Advice to younger self: things aren’t as hard as they look—starting is easier than finishing/exiting
- 5:32 – 7:35
From Nicaragua to the American dream: grit, immigration, and manufactured adversity
Elias recounts growing up amid revolution and instability in Nicaragua, becoming effectively homeless, and moving to the US through family sacrifice. He argues that hardship builds grit, and that immigrants often thrive in startups because they’ve already endured real adversity.
- •Revolution, displacement, and a green card timing that forced a life reset
- •The US is still a uniquely high-opportunity environment in his view
- •Cultural contrast: what people complain about vs. true hardship
- •How he tries to “manufacture adversity” for kids growing up with privilege
- 7:35 – 11:34
Finding tech early and learning inside IBM’s ‘advanced internet’ era
He describes his first exposure to computers and how it planted the seed for building. At IBM, he worked on early web applications and learned both the power of visionary teams and the limits of scale inside giant organizations.
- •Early hands-on computing: WordPerfect, Lotus 1-2-3, curiosity-driven learning
- •IBM’s advanced group: building browser-based apps in the late 90s
- •Big-company constraints: coordination overhead vs. speed
- •Setting up the later thesis: ‘elephants can’t dance’
- 11:34 – 15:57
Why incumbents feel fast but aren’t: ‘elephants can’t dance’ in the AI era
Harry challenges Elias on whether large incumbents are moving faster than ever, and Elias pushes back hard: distribution isn’t the same as speed. They debate Google and Microsoft specifically, and why big organizations struggle to ship truly great new products quickly.
- •Distinguishing distribution advantage from genuine execution speed
- •Skepticism of incumbent AI products: lots of announcements, uneven utility
- •Microsoft’s OpenAI deal as a brilliant strategic move, but product adoption challenges remain
- •Core constraint: too many people, approvals, and risk-avoidant incentives
- 15:57 – 20:17
The ‘operator’ problem: why ChatGPT isn’t the end state and agents will reshape software
Elias lays out a mental model: traditional software assumes a human operator, like a car needing a driver. The future is ‘driverless’—agentic systems that execute workflows with minimal human involvement, making today’s interfaces a transitional stage.
- •Software built for operators vs. software built for autonomy
- •ChatGPT as a major step function but not the final form
- •Agentic workflows as the real destination: systems doing the work, not assisting the work
- •Parallels to autonomous driving levels and the gradual path to full autonomy
- 20:17 – 22:20
Small teams, big revenue: AI as labor replacement and the coming org-design reset
They explore how AI shifts cost structure from headcount to compute, enabling much smaller companies to achieve massive scale. This leads into a discussion of what that means for venture financing, equity dilution, and a possible move toward debt in later stages.
- •AI substitutes compute for labor, reducing the need for large support and ops teams
- •‘Smaller is better’ becomes feasible at billion-dollar revenue levels
- •Enterprise bloat: hiring, training, management layers, and CS scaling pain
- •Implications for VC: less equity burn on headcount; potential shift to debt for growth
- 22:20 – 26:50
Enterprise adoption bottlenecks: org politics, handoffs, and CEO accountability
Elias describes selling into enterprises and the predictable failure mode: enthusiastic CEOs delegate the decision downward to teams that resist change—especially if AI threatens their roles. He argues real adoption requires top-down ownership, and expects a slower, hype-heavy transition.
- •The ‘pushdown’ problem: CEO says yes, middle layers stall or block
- •Resistance is strongest among the people most disrupted by automation
- •AI services/handholding as a bridge until products become truly agentic
- •Longer timeline belief: slow real-world adoption and incomplete solutions today
- 26:50 – 33:52
How Elias hires “rockstars”: high agency, speed, obsession—and an intense interview style
Switching to talent, Elias explains what he looks for: high agency, grit, execution speed, and customer obsession. He describes a highly confrontational, proof-driven interview style designed to stress-test candidates in real time.
- •Hiring criteria: IQ + grit + speed + obsession + ‘high agency’
- •Prefers extroverted, customer-facing engineers; intensity filters for fit
- •Proof over talk: “open your computer, show your Git, show your code”
- •The interview is the test: if you can’t handle the pressure, you won’t thrive there
- 33:52 – 42:03
Process upgrades: contracting, pushback, and ‘hire slow, fire fast’
Elias contrasts his earlier growth-era hiring (fast, persuasive, sometimes too accommodating) with his current approach at Agency. Now he emphasizes trials, resistance-testing, and decisive exits when signals are negative—often within 1–2 weeks.
- •Old approach: ‘numbers game’ hiring and over-convincing candidates
- •New approach: reject/push back to see who persists and performs
- •Contract-first trials to evaluate real output, not interview performance
- •Decisiveness: most founders know quickly but ignore signals due to ego
- 42:03 – 47:08
Red flags, titles, and incentives: what founders get wrong about teams and motivation
They discuss recurring hiring mistakes—job hopping, superficial title signals, and story-heavy candidates. Elias argues titles inflate ego and bureaucracy, and that companies must rethink rewards and incentives as orgs become smaller and more profit-focused.
- •Bouncing careers and shallow narratives as warning signs; probe for real work
- •Titles as ego-fuel: pushes for alternative rewards (profit/revenue sharing, equity)
- •Founders hire ‘good enough’ under pressure—then regret the drag for years
- •Shift in mindset: durability, profitability, and legacy matter more post-ZIRP
- 47:08 – 52:12
‘Drift was my biggest failure’: too many people, product timing, and the reality of selling
Elias calls Drift a personal failure despite the $1.2B outcome, describing how scaling headcount and losing product momentum undermined durability. He walks through the sale decision, the ego hit of no longer mattering in ‘your’ company, and the emotional aftermath of exit.
- •Failure framing: not building an enduring, customer-delighting product foundation
- •Root causes: bloat, inefficiency, execution speed, and timing relative to LLMs
- •Sale context: 2021 valuations, ego comparisons, and choosing realism over pride
- •Post-exit identity dip: ‘being nobody’ and the psychological crash after selling
- 52:12 – 1:07:27
Working parents, kids early, and marriage: redefining ‘peak performance’ and toughness
Elias argues for having kids earlier, citing biology, societal demographics, and the long-term payoff of growing alongside your children. He and Harry debate whether family harms founder performance, and Elias shares lessons on marriage, parenting styles, love languages, and balancing affirmation with toughness.
- •Pro-kids-early thesis: fertility realities, IVF difficulty, and demographic decline
- •Doing both (startup + parenting) is possible but “ugly” and exhausting
- •Marriage advice: reduce selfishness, ignore social-media illusions, prioritize others
- •Parenting paradox: be loving/affirming while still manufacturing adversity
- 1:07:27 – 1:11:34
Quickfire and closing optimism: directness, bets, fears, and rebuilding craftsmanship
In quickfire, Elias emphasizes directness, skepticism about investing prowess, and concerns about extremism and ignorance. He closes with optimism inspired by Egypt’s history—hoping AI enables better products, deeper customer relationships, and more time for culture and family.
- •Belief: directness beats indirectness in action and communication
- •Compute thesis: picks NVIDIA; calls out Salesforce as a short
- •If he couldn’t fail: he’d educate more people globally
- •Optimism: return to craftsmanship, better business practices, more art/culture/family time
