Move Fast, Break Nothing: Why Government Doesn’t Run on Startup Logic

Sahil Lavingia joined Elon Musk's government efficiency team with Silicon Valley confidence. Fifty-five days later, he got fired for telling reporters what he found inside. His account reveals who really ran DOGE—and why the VA surprised him.

Move Fast, Break Nothing: Why Government Doesn’t Run on Startup Logic
Symbol Photo / AI created image

💡 TL;DR - The 30 Seconds Version

🔥 Gumroad founder Sahil Lavingia got fired from DOGE after 55 days for talking to reporters about his experience inside Trump's efficiency initiative.

📊 Elon Musk departed DOGE after hitting his 130-day legal limit, leaving Steve Davis as the real operational leader coordinating through encrypted Signal messages.

💰 DOGE claims $170 billion in savings but experts estimate $80 billion is more accurate, while federal spending actually increased 8.7% since January.

🏛️ Government layoff rules from 1944 shocked Silicon Valley types by prioritizing seniority over performance, cutting newest hires first regardless of productivity.

🎯 Lavingia found the VA less broken than expected, saving real money like ending a $380,000 monthly website contract now done by one engineer.

🚀 Tesla sales collapsed 50% in Europe amid boycotts tied to Musk's political role, showing the business cost of his DOGE involvement.

Sahil Lavingia thought he'd help revolutionize government. Instead, he learned that bureaucracy exists for reasons beyond incompetence.

The Gumroad founder joined Trump's Department of Government Efficiency in March as an unpaid software engineer at the Department of Veterans Affairs. Fifty-five days later, he got fired for talking to a reporter. His account reveals the gap between DOGE's revolutionary promises and bureaucratic reality.

Lavingia arrived with Silicon Valley confidence and Bernie Sanders idealism. He wanted to write code that helped people at scale. The VA seemed perfect—a $350 billion agency with 473,000 employees that surely needed tech disruption.

His first assignment was reviewing contracts for cancellation. The VA had over 90,000 active contracts covering everything from IT support to landscaping. Lavingia proposed a simple solution: use AI to scan contract PDFs and flag wasteful spending. Build a dashboard for easy review. Classic startup thinking.

The Reality of Government Layoffs

Then he learned about Reduction in Force rules.

Government layoffs follow strict protocols dating to the Veterans' Preference Act of 1944. Tenure matters most—newest hires go first. Veterans get protection over non-veterans. Seniority beats performance. Anyone promoted within two years counts as probationary and faces the axe.

This wasn't the bloated middle management purge he expected. It was algorithmic, based on service records rather than productivity. Performance ratings only broke ties at the very end.

"These reduction-in-force rules surprised me and many others," Lavingia wrote. "Unlike private industry layoffs that target middle management bloat and low performers, the government cuts its newest people first, regardless of performance."

What He Found at the VA

The revelation continued. The VA wasn't as broken as he assumed. It had already consolidated HR systems. Engineers were building ambitious projects to cut veterans' benefits processing from 133 days to under a week. The world's first electronic health record system, VistA, came from VA employees 40 years ago.

His government laptop couldn't run Git or Python due to security policies. But the VA's CTO recommended getting a proper MacBook for engineering work. Problems had solutions—they just moved slower than startup speed.

DOGE's Chaotic Internal Structure

Meanwhile, DOGE itself proved surprisingly disorganized. Lavingia expected a cohesive team culture. Instead, he found scattered engineers starting from scratch with no central playbook or knowledge sharing. It felt more like having McKinsey consultants embedded across agencies than a revolutionary force.

The real power structure became clear. Elon Musk appeared at occasional meetings. Steve Davis, a longtime Musk lieutenant from X and the Boring Company, coordinated day-to-day operations through encrypted Signal messages. Anthony Armstrong and Baris Akis rounded out the leadership trio.

"Steven was the only person who was across everything," Lavingia told WIRED. "Steven is basically like a chief of staff or body man when Elon was there."

Davis would message priorities to DOGE team leads at different agencies. At the VA, he instructed the team to focus on reviewing contracts for cancellation. He'd check in periodically but rarely responded to updates.

Limited Power, Maximum Blame

DOGE had no direct authority to fire anyone or cancel contracts. Those decisions came from Trump-appointed agency heads who let DOGE take blame for unpopular cuts. "The real decisions came from the agency heads appointed by President Trump, who were wise to let DOGE act as the 'fall guy' for unpopular decisions," Lavingia observed.

This mirrors Musk's recent complaints to the Washington Post about DOGE becoming Washington's "whipping boy."

Building Tools That Actually Worked

Despite constraints, Lavingia built useful tools. He created contract analysis software using large language models to flag potential waste. He deployed interactive org chart visualizations for the massive VA workforce. He modernized the agency's internal ChatGPT system (seriously called VAGPT) with a mobile-responsive interface built on NextJS and Tailwind.

His work had real impact. According to VA leadership, one contractor was charging $380,000 monthly for minor website modifications. That contract wasn't renewed. The same work now gets done by one internal engineer spending 10 hours per week.

But Lavingia never got approval to ship anything major that would directly help veterans. His prototypes for improving disability claims filing and automating claims processing stayed in development. The bureaucratic approval process moved too slowly for startup-style iteration.

He also worked on projects that revealed DOGE's political priorities. His open-source code includes a tool that scanned internal PDFs for terms "related to DEI, gender identity, COVID policies, climate initiatives, WHO partnerships."

The Abrupt End

The end came abruptly. After discussing his DOGE experience in a Fast Company interview, his access got revoked without warning. No explanation. Just locked out.

His departure coincided with broader DOGE upheaval. Musk reached his 130-day limit as a Special Government Employee and officially left this week. Davis and Nicole Hollander also appear to be departing. With the operational leader gone, DOGE's future direction remains unclear.

The timing hurt Musk's other ventures. Tesla sales collapsed 50% in Europe and 87% in Quebec amid boycotts and protests tied to his political involvement. The latest SpaceX Starship test ended with a "rapid unscheduled disassembly"—engineer speak for explosion.

The Disputed Results

DOGE's accomplishments remain disputed. The official website claims $170 billion in taxpayer savings through canceled contracts and layoffs. But reports suggest the "Wall of Receipts" contains misleading figures. A conservative think tank fellow estimates the real number around $80 billion.

More tellingly, federal spending actually increased 8.7% since Trump took office, largely due to Social Security, Medicare, and debt payments. The full impact of DOGE cuts may not show up in immediate figures, given upfront costs of ending programs and firing workers.

Lavingia's experience captures the classic tension between disruption and stability. Silicon Valley assumes existing systems are broken and need replacing. Government assumes existing systems serve purposes beyond efficiency.

Both can be right. The VA was paying $380,000 monthly for work worth a few thousand. But it also had working systems, dedicated employees, and legitimate constraints that prevented startup-style "move fast and break things" approaches.

"The government works," Lavingia concluded after his deep dive. "It's not as inefficient as I was expecting, to be honest. I was hoping for more easy wins."

Why this matters:

  • DOGE's internal dysfunction and limited authority explain why bold promises of government transformation fell short—it turns out you can't run government like a startup.
  • Lavingia's firing for speaking honestly about his experience shows how Trump's efficiency crusade prioritizes loyalty over the transparency it claims to champion.

❓ Frequently Asked Questions

Q: Why could Elon Musk only serve 130 days in DOGE?

A: Federal law limits Special Government Employees to 130 days of service per year. Musk hit this limit on May 30, 2025, forcing his departure despite DOGE's mandate running until July 2026.

Q: What's VAGPT and why does the VA have its own ChatGPT?

A: VAGPT is the Department of Veterans Affairs' custom version of ChatGPT. Lavingia modernized its interface to be mobile-responsive. Federal agencies often create internal AI tools due to security restrictions on using public versions.

Q: How much did DOGE actually save taxpayers?

A: DOGE claims $170 billion in savings, but experts dispute this figure. A conservative think tank estimates $80 billion is more accurate. Meanwhile, federal spending increased 8.7% since Trump took office.

Q: Who really ran DOGE day-to-day operations?

A: Steve Davis, Musk's longtime lieutenant from X and the Boring Company, coordinated daily operations through encrypted Signal messages. Lavingia called him "the only person who was across everything."

Q: Why did government layoff rules surprise Silicon Valley types?

A: The Veterans' Preference Act of 1944 requires firing newest hires first, regardless of performance. Veterans get protection over non-veterans. Seniority beats skill. This contrasts sharply with private sector layoffs targeting underperformers.

Great! You’ve successfully signed up.

Welcome back! You've successfully signed in.

You've successfully subscribed to implicator.ai.

Success! Check your email for magic link to sign-in.

Success! Your billing info has been updated.

Your billing was not updated.