Summary: The Department of Government Efficiency (DOGE), orchestrated by Elon Musk, has begun embedding private-sector operatives into the U.S. Department of Veterans Affairs (VA) with a clear mission: rewrite the agency’s technological DNA using artificial intelligence, automation, and radical cost-cutting. But beneath the surface of tech rhetoric and efficiency slogans lies a high-stakes gamble. This post unpacks the individuals involved, the tools being introduced, and the existential risk to the veterans served by America’s largest integrated healthcare system—all under the guise of “digitization.”
The Technocrats Arrive: DOGE’s Silicon Valley Foot Soldiers in a Federal Behemoth
On March 25, VA tech teams noticed something strange in their internal systems—an unknown name, Sahil Lavingia, accessing core development repositories on GitHub. Lavingia isn’t a civil servant. He’s the founder of Gumroad, a digital storefront platform known for layoffs, radical automation, and minimal staffing. Without any public-sector or defense medical infrastructure background, he had somehow landed a privileged position: adviser to the VA Chief of Staff, Christopher Syrek.
That one digital fingerprint set off alarm bells, and for good reason. Lavingia’s involvement signaled the operational start of DOGE’s embedded strategy: place startup insiders inside government agencies to redesign them by private-sector logic. In this case, it’s the VA—an institution that delivers healthcare, financial services, housing, and education to 10 million veterans.
Open Source Intrusion: When AI Starts Writing Federal Code
Lavingia made his intentions known early: make the VA digital-first, AI-driven, and “lean” in Silicon Valley terms. Reports from VA insiders reveal that he attempted to introduce an AI tool called OpenHands directly into the GitHub development environment—software that, while promising, hadn’t passed federal security protocols. What’s at stake? Sensitive data including Social Security numbers, disability records, and home addresses.
How do we justify integrating generative code tools before evaluating their ability to safeguard veteran medical histories, financial identities, and system integrity?
VA staff say they were pressured to explain “why AI couldn’t do the work” every time a development project came up. In short: DOGE’s approach assumes AI automation is the default, and human discretion the exception.
Privileged Access Without Accountability
One of the more worrying details is Lavingia’s access level within the VA. Multiple sources confirmed he had a “zero account”—a privileged user type that can bypass many regular security restrictions. Pair that with an AI writing code inside platforms accessing federally protected health information (PHI), and you have a recipe for mass exposure with minimal oversight.
Who authorized this access? What assessment protocols were followed? If none, what precedent is being set for other agencies?
Who Else Is Inside? The DOGE Operational Roster at VA
Lavingia isn’t working alone. Other known DOGE operatives embedded at the VA include:
- Cary Volpert – formerly involved in social engagement for seniors, now senior adviser to the Chief of Staff.
- Christopher Roussos – ex-CEO of 24 Hour Fitness, and a veteran of corporate healthcare venture AllerVie Health.
- Justin Fulcher – founder of a failed telehealth startup.
- Payton Rehling and Jon Koval – former employees at Valor Equity Partners, a Musk-backed investment firm.
Is this a strategy to disrupt bureaucracy or an ideological takeover of government operations? Many VA workers believe it’s dangerously closer to the latter, with staff morale and mission clarity becoming early casualties of Musk’s efficiency experiment.
The Fallout: From Paper Forms to Recordings Without Consent
Lavingia didn’t just target backend changes. He also suggested sweeping administrative replacements such as eliminating all paper forms—vital tools for elderly or visually impaired veterans—and automating front-end processes that hypothetically could be abused to submit fraudulent claims. One suggestion he floated involved extracting and auto-filling data using Social Security numbers or “other unique identifiers,” even for users not logged into the system.
Who decides how much friction is too much when the loss of friction means easy data exploitation?
Even meeting norms have shifted under DOGE’s influence. During a Microsoft Teams meeting, VA staff were startled by an unannounced recording started halfway through a high-level conversation. What message does it send when consent and procedures are treated as optional?
The Price of Disruption: Real People, Real Risk
Veteran-facing programs aren’t just service lines—they’re safety nets. The VA’s sprawling infrastructure covers healthcare, rehab, housing, education, and financial security. Replacing trained contractors and administrators with half-tested AI tools—without rigorous risk assessments—puts millions at risk of being digitally locked out. Or worse, of having their lives disrupted by bugs, outages, or breaches.
Replacing Drupal, a CMS that on-site VA staff use daily, with raw code updates would sideline non-technical employees and compromise site functionality. After Lavingia’s recommendation, sources say the VA canceled vendor contracts serving the CMS system—effectively severing everyday agency staff from basic content updates.
Do these technologists understand the operational crush of a rural VA clinic trying to update holiday hours when the system now requires code commits instead of a login?
Silicon Valley’s Playbook Collides with Public Service
The cultural clash couldn’t be starker. As one tech staffer said, “They think because they built a commerce site or fitness chain, they can design federal programs for disabled Vietnam vets.” But the VA is not a startup. It’s not a consumer app. It’s not a sandbox for disruption theater.
And yet, the newcomers garner fear and obedience—not because of their insight but because of their connection to Elon Musk, Donald Trump, and the wider DOGE directive. Inside the agency, questions are now treated like resistance. Staff monitor not only the codebase but also their own heads—self-censoring, doubting, and wondering what will be broken next.
Is this about veterans, or about proving that the Silicon Valley elite can do government better than government?
The Larger Stakes: Efficiency vs. Human Dignity
Everyone agrees the VA should improve processes and technology. But the means matter. Veterans are not edge cases in an A/B test. They are individuals who entrusted their bodies and minds to the nation’s defense. Services they rely on—health procedures, disability benefits, housing, GI Bill programs—cannot afford outages or creative rewrites by people who’ve never had to file a claim, treat a case of PTSD, or navigate benefits for a fallen comrade’s widow.
Ethics and accountability must remain above innovation theater. The real question: if DOGE’s AI-driven redesign causes a benefits disruption that costs lives or stability, who will take responsibility?
Bold moves can save institutions, but blind faith in automation—without empathy, process, or checks—won’t.
#VeteranCare #FederalAI #TechInGovernment #VADigital #DataSecurity #EthicsInAI #DogeWatch #MuskTakeover #StartupMindset #PublicSectorRisks
Featured Image courtesy of Unsplash and Ales Nesetril (Im7lZjxeLhg)
