Gutenberg Lab — Morning Paper

AI Is Breaking the 13-Year-Old "Give Away Your LEGOs" Career Framework — Grief, Loneliness, and Burnout Sweep Through Tech

Featuring: Lenny Rachitsky × Molly Graham (former Google/Facebook exec, leadership advisor)

Bottom line:

Delegating to AI is psychologically different from delegating to humans because you cannot eliminate oversight — the mental burden stays with you, which explains why burnout jumped from 44% to 55% even as productivity rose.

3-Line Summary

Molly Graham's 13-year-old "give away your LEGOs" career advice (delegate work as companies scale) faces its first major revision in the AI era. When you give work to humans, you can "chuck it at their face and run"—fully transferring ownership. With AI, you still own the final product and the mental tax never goes to zero. Graham and Rachitsky examine widespread grief, loneliness, and burnout in tech, arguing that while change is inevitable, leaders must create space for emotional processing and help teams distinguish between work that should be automated and uniquely human contributions that must be protected.

Key Points

1. Burnout jumped 11 points in one year — but happiness also hit all-time highs

Rachitsky's survey shows burnout rose from 44% (2024) to 55% (2025), yet half of respondents report feeling the happiest they've ever been in their careers. Strongest correlation with happiness: people who feel AI has amplified them. Designers are the least happy—everyone's moving faster, but design requires feedback, alignment, and thinking that can't be iterated with agents.

2. "Everyone is a manager now" — managing robots requires the same skills as managing humans

Graham: "Delegating to robots is delegating to team members. It requires oversight, context, correcting work, coaching. Literally everyone in the world is a manager now." Many people actively chose not to become managers and don't enjoy delegation. Happiest managers max out at 10-12 direct reports; counting robots is concerning for people's psychology.

3. Code rewrite rate up 8x — productivity ≠ efficiency

Graham cites an engineering report: AI made engineers more productive, but lines of code requiring rewrites increased 8x. Security incidents also rose. The industry focused on "token-maxing" and productivity metrics, but missed the efficiency question: "Does it move the conversation forward?"

Editorial Perspective

Graham's first caveat in 13 years—"some LEGOs shouldn't be given away"—captures the fundamental difference between this platform shift and previous ones. The birth and rise of the internet required developers to learn new distribution channels, but the work itself stayed the same. The AI transition is redefining the work itself, and unless humans protect "things requiring judgment," "things requiring trust," and "things where you don't know the definition of good," productivity numbers will rise while output quality collapses. Zomorodi's reframe—"your job won't go away, it'll just look different every six years"—is the antidote to the fear narrative.

Source: Lenny's Podcast "The grief, loneliness, and burnout sweeping through the tech industry right now | Molly Graham" (September 27, 2026)

https://www.lennysnewsletter.com/p/the-grief-loneliness-and-burnout

AI Disclosure: Produced with AI assistance; facts and analysis reviewed by our editorial team.

• --

The Super Shoe Revolution That Broke the 2-Hour Marathon Barrier — A Record Scientists Predicted for 2075 Arrived Half a Century Early

Speakers

Lauren Feiner (Senior Policy Reporter, The Verge) × Alex Hutchinson (Sports Science Journalist, Author of *Endure*)

Bottom line

Breaking the two-hour marathon required the convergence of carbon-plated shoes (4% efficiency gain), doubled carbohydrate intake (120g/hour via hydrogel technology), and generational athletic talent — we are witnessing technology rewrite the definition of human limits in real time.

Three-line summary

In the 2026 London Marathon, Sebastian Soway and Joomov Kajalcha became the first to break two hours in an official race, half a century earlier than Hutchinson's 2014 prediction of 2075. The primary driver was "super shoes" — thick resilient foam with embedded carbon fiber plates — starting with Nike's Vaporfly 4%, which improved efficiency by 4%. Foam energy return has since evolved from 70% to 90%, and the innovation has not plateaued because scientists still do not fully understand how the shoes work.

Three key points

1. The mechanism remains unexplained after a decade: Nike's Vaporfly 4% demonstrated a 4% efficiency gain, but scientists still disagree on how the thick sole, resilient foam, and carbon plate interact. Removing any component degrades performance, but why the combination works remains contested — empirical optimization is outpacing theoretical understanding

2. The nutrition revolution shattered a physiological ceiling: The old limit was 60g of carbohydrates per hour; Soway and Kajalcha consumed approximately 120g/hour (equivalent to six plates of pasta) using hydrogel encapsulation technology that hides carbs from stomach sensors and releases them in the intestine. Hutchinson notes this "would have been dismissed as absolutely insane five years ago"

3. A $500 disposable shoe sold out in five seconds: The Adidas Pro Evo 3 (worn by both runners) is the first sub-100g competition shoe, "optimized to fall apart after 27 miles." Five hundred to one thousand pairs sold for $500 and immediately resold for approximately $2,000. The market will pay extreme premiums for marginal gains when durability is sacrificed for weight

Editorial perspective

The speed at which technology rewrites the definition of "impossible" betrays expert predictions by half-century margins. Hutchinson spent months researching for his 2014 prediction of 2075, yet missed how incremental materials science improvements could produce step-change outcomes. Super shoes face criticism as "springs on your feet," but all shoes are springs — the difference is efficiency. The real problem is that efficiency created a collective action problem: Boston Marathon qualifying times were lowered by five minutes twice, making it impossible to "unilaterally disarm." Whether you buy a $500 shoe that lasts two races depends on what you believe running should be about.

Source: The Vergecast, "The tech that made a 2-hour marathon possible" (September 8, 2026)

https://www.theverge.com/the-vergecast

Spoken source: The Vergecast, "The tech that made a 2-hour marathon possible" (September 8, 2026). Guest: Alex Hutchinson

AI Disclosure: Produced with AI assistance; facts and analysis reviewed by our editorial team.

• --

Legal AI firm Harvey bets on trust over process to scale to 2,000 while keeping experimental culture

Featuring

Katie Kirsch (a16z) × Maggie Landers (VP of Talent at Harvey)

Bottom line

In AI-era scaling, the dominant choice is whether you kill speed by adding process or preserve it by maintaining trust and removing friction.

3-Line Summary

Legal AI company Harvey scaled from 340 to nearly 2,000 employees in one year while its valuation jumped from $3 billion to $11 billion. With over 70% of the workforce having less than one year tenure, the company rejects the typical scaling playbook of adding hierarchy and process, instead betting on founder accessibility, employee trust, and a culture that permits failure. VP of Talent Maggie Landers describes growing the recruiting team from 16 to 80 people in one year to support hiring across 400+ open roles.

3 Key Points

1. Valuation trajectory and structural knowledge vacuum: Valuation rose from $3 billion (February 2025) to $11 billion (March 2026). Headcount increased by over 1,000 in one year, reaching approximately 2,000 by end of 2026. More than 70% of employees started after January 1, 2026, creating a situation where recruiters, hiring managers, and leaders are often all new simultaneously—forcing real-time process invention rather than institutional knowledge transfer.

2. Overcoming the "A+ student syndrome": Most Harvey employees were previously successful high achievers who must actively shed perfectionism. Maggie notes "you're gonna fail more than you're necessarily gonna get it right" at Harvey. The company's value of "don't take the square root of the weather" (founder Winston Shi's phrase) means prioritizing action and learning over overanalysis—a decision-making philosophy that accepts higher error rates for speed.

3. Trust as competitive moat, not soft culture: Winston recently said in an AMA that "trust" is what he wants to preserve most as Harvey grows. Maggie frames this as operational, not cultural: "The trust that we have for our employees and frankly, the agency that we give folks to just run and execute without a lot of the policing and without a lot of the friction is what continues to allow us to hire the amazing folks that we do and move as quickly as we do." While most companies add process at scale, Harvey bets that removing friction is the scaling strategy.

Editorial Perspective

Transforming the legal industry requires different talent than building models or infrastructure. Harvey distributes legal expertise across all functions—applied legal researchers in product, legal engineers in sales—choosing to stay close to customers rather than abstract the domain away. This is a structural bet that technical superiority alone cannot win in application-layer AI; you need domain expertise embedded throughout, not concentrated in one team. Whether Harvey can reach 2,000 people while maintaining trust and speed is an experiment in the scalability limits of AI application companies themselves.

Source: a16z "Building a Team at AI Speed | Harvey's Maggie Landers" (September 27, 2026)

https://a16z.simplecast.com/episodes/building-a-team-at-ai-speed-harveys-maggie-landers-1mxeFqtp

AI Disclosure: Produced with AI assistance; facts and analysis reviewed by our editorial team.

• --