How the Best CEOs Delegate

Jonathan's formative experience working in the West Wing next to the president's executive assistants set an impossibly high bar for what EA partnerships could achieve. This led him to build Thumbtack with a half-dozen assistants, discovering that more leverage creates more ambition in a compounding cycle.

• •Working alongside presidential EAs in the West Wing established world-class standards for delegation
• •Started with one assistant in the Philippines for basic tasks, then scaled to six specialized assistants
• •Key insight: More leverage generates more ambition, not the other way around
• •Democratization of elite assistance: what cost $500k for Marc Andreessen now costs $3k/month with Athena

Jonathan explains the evolving relationship between human and AI assistants using the self-driving car framework. AI capabilities will increase gradually, starting with $20/month ChatGPT for those on a budget, scaling up to full human assistant teams for those with resources.

• •If you don't have an assistant, you are the assistant - start delegating even with just ChatGPT at $20/month
• •Prompt engineering is fundamentally delegation - master this before hiring humans
• •AI will become increasingly autonomous like Tesla's assisted driving, but humans provide the UX layer
• •Budget ladder: $20/month ChatGPT → $3k/month Athena → $100k+ in-person assistant → full suite of specialists

The cardinal sin of delegation is thinking 'it will be faster to do it myself.' Jonathan shares how to move from task-based delegation to exporting your internal algorithms, using examples like planning founder dinners that led to meeting his wife.

• •Delegate by algorithm, not by task - export your internal decision-making process and preferences
• •Example: Instead of 'plan a dinner,' specify '6-8 people, similar funding stage, similar employee count'
• •Voice delegation is 2-3x faster than typing - use it between meetings to prevent work accumulation
• •The activation energy of teaching someone is worth it for long-term compounding leverage
• •Specialize assistants by domain: work, home, kids, travel, finances for maximum efficiency

Jonathan emphasizes that switching assistants every 6-12 months destroys the compounding effect. His assistant Marnie has worked with him for a decade, helping with everything from founder dinners to wedding planning to raising kids.

• •Common mistake: switching assistants frequently prevents compounding leverage
• •Start with email/calendar access, earn trust over time before giving bank account access
• •Best super-delegators at Athena all use voice delegation constantly throughout the day
• •Future vision: AI watches your screen, automatically identifies delegatable tasks without vocalization
• •Machine-generated delegations already working internally - majority of tasks pulled automatically

Jonathan explains why the Philippines works well for executive assistants - strong affinity to American culture, caretaking culture, and strong work ethic. He addresses the mindset shift needed to see delegation as creating opportunities, not indulgence.

• •Philippines selected for cultural affinity to US, caretaking culture (many nurses), and work ethic
• •Reframe delegation guilt: you're creating life-changing jobs for people excited to help you succeed
• •First project: have assistant save money through subscription audits to pay for their own salary
• •Scale from admin (paying bills) to coordination (PM between accountants/lawyers) to full financial management
• •Assistants provide accountability and coaching - some clients work out live with their assistants

Jonathan and his wife Catherine apply quarterly business review rigor to their relationship, inspired by Clayton Christensen's 'How You Measure Your Life.' He emphasizes identifying the one power law goal each quarter that's worth more than everything else combined.

• •Brian Johnson breaks chains of biology; Jonathan wants to break chains of time - the most valuable asset
• •Quarterly relationship reviews with survey, SWOT analysis to prioritize what matters most
• •Power law in everything: identify the ONE thing each quarter worth more than all else combined
• •Long goal lists enable procrastination on the most important foundational thing (partner, business, health)
• •Calendar audit: if last month doesn't reflect your highest goals, you're not doing it right

The more senior an executive, the better they are at interviewing, making the interview itself less useful. Jonathan relies heavily on references, asking for 360 reviews, and sourcing through high-bar networks rather than recruiters.

• •Senior executives are great at interviewing by definition - discount the interview at C-suite level
• •Ask candidates for their 360 reviews from previous companies to see honest strengths/weaknesses
• •Reference check technique: 'I'm probably going to hire them, now tell me what they're bad at'
• •Best hiring method: ask 3 people with highest standards for 2-3 best people, then recruit from that list
• •Project-based interviews: have candidates solve actual problems they'll face in the role

Jonathan shares the story of Google giving Thumbtack the 'death penalty' overnight, eliminating all traffic and revenue. He explains why you can't share everything during crisis, and reflects on having four cofounders (three non-technical) - not recommended but it worked.

• •Google crisis: woke up to zero traffic, zero revenue - had to handle before telling new employee class
• •Default to transparency, but during crisis get handle on situation before company-wide communication
• •Michael Moritz insight: 'There's only one founder' - eventually only makes sense to have one person on top
• •Pick cofounders like marriage partners - you'll have relationship issues without makeup sex
• •Four cofounders (three non-technical) is 'idiotic way to start tech business' but worked through specialization

EAs are typically administrative (defensive), while chiefs of staff are offensive - capable of becoming founders themselves. Jonathan recommends going deep with EAs for a decade of compounding, but accepting 2-year tours with high-slope chiefs of staff.

• •EA = administrative, taking things off your plate; Chief of Staff = offensive, can take meetings for you
• •Chiefs of staff have founder-level capability, do tours of duty to learn before starting their own company
• •EA strategy: find someone for a decade of compounding intimacy and trust
• •Chief of staff strategy: pure slope, accept 2-year tenure, invest in their next company
• •Examples: Lonsdale, Thiel, Bezos build alumni networks of former chiefs of staff who become founders

Jonathan challenges the assumption that wealthy people can afford big teams. The data shows the opposite: as people get more leverage, their ambition increases. He shares examples from billionaire hedge fund managers with 40-person teams to Elon's claimed lack of assistants.

• •Counterintuitive insight: leverage creates ambition, not the other way around
• •Hard to have big ambition when weighed down by DMV forms and passport renewals
• •Billionaire hedge fund manager: 8 Princeton-grad EAs, will spend 1 day to save principal 1 minute
• •Elon claims to be his own assistant (Bill Gates couldn't believe it) - but he's an outlier
• •Most successful people hold remarkably few grudges - desire to win trumps past conflicts

Building a company is 'shock therapy' - you eventually get used to bathing in existential fear. Jonathan emphasizes that Thumbtack had multiple near-death experiences, but staying in the game long enough is how you eventually win.

• •Founder psychology is like cold plunge - shock therapy until you're comfortable with existential fear
• •Thumbtack had many near-death moments but survived by staying in the game long enough to win
• •Degrees of panic: mean internet comments → employee departures → 6 months runway → lawsuits
• •Eventually success means reviewing a list of lawsuits quarterly - it becomes part of business
• •Cofounders are critical for having someone to be totally open with during the ups and downs

Athena started fully bootstrapped and human-only, scaling to 1,000 people with no outside capital. Now building the future where humans provide UX and empathy while AI handles increasingly complex mechanical tasks, with humans training the models through their work.

• •Athena scaled to 1,000 people bootstrapped before making big AI play to avoid being railroaded
• •Vision: best human + best machine in unified product, with machine capabilities increasing over time
• •Humans focus on project management, high-level work, empathy; AI handles rote mechanical tasks
• •Every task a human does trains the AI models - like Tesla drivers training self-driving
• •Hiring bar: 50,000 applicants per month, 1 in 300 hired after extensive testing
• •If hiring yourself on Upwork, interview 50+ people, not just 10, to match quality