A breach at a16z-backed Doublespeed reveals a physical army of 1,100 phones flooding TikTok with undisclosed ads and AI-generated “slop.”
- The “Legitimate” Phone Farm: Doublespeed, backed by venture capital giant Andreessen Horowitz, has industrialized spam tactics by using thousands of physical phones to mimic human behavior and evade platform detection.
- A Massive Breach: A hacker successfully seized control of the startup’s 1,100-device fleet, exposing a network of covert AI influencers promoting products on TikTok without proper ad disclosures.
- The Era of Platform Decay: The incident underscores a growing crisis where VC-funded “innovation” fuels the degradation of online spaces, violating terms of service while platforms struggle to moderate the influx of artificial noise.
We are all familiar with the modern internet’s background radiation: those obnoxious social media accounts that flood comment sections and inboxes with nonsensical spam. Historically, this has been the domain of financial criminals and hackers operating in the shadows. However, a recent revelation has exposed a new reality where this tactic isn’t just a scam—it is a legitimate business model backed by one of Silicon Valley’s most powerful venture capital funds.
Enter Doublespeed, a fledgling startup that bills itself as a “bulk content creation” service. In reality, it operates a massive “phone farm”—a tactic traditionally associated with ad fraud and fake engagement. Backed by a $1 million cash injection from a16z (Andreessen Horowitz), the firm founded by Marc Andreessen and Ben Horowitz, Doublespeed has effectively gentrified the click farm. For fees ranging between $1,500 and $7,500 a month, clients can access this mechanical army to “orchestrate actions on thousands of social accounts.”
The Hack That Revealed the Machine
The true scale and nature of Doublespeed’s operations were forcibly brought into the light when an anonymous hacker gained access to the company’s infrastructure. The breach allowed the intruder to take control of 1,100 physical smartphones—the hardware powering the company’s “instrumented human action.”
This phrase is industry jargon for a system where bots control physical devices to tap, swipe, and scroll, mimicking natural user behavior to fool algorithms into thinking the engagement is human. The hack revealed that these devices were powering covert, AI-generated ads on TikTok. Perhaps more troubling was the discovery that the AI-generated accounts were promoting products often without the legally required disclosure that the content was an advertisement.
The vulnerability of the system adds a layer of irony to the company’s technical claims. In a post on X (formerly Twitter), Doublespeed co-founder Zuhair Lakhani once boasted about the heavy reliance on artificial intelligence for development, stating, “Claude code is truly our third cofounder.” It appears that this “cofounder” may have left the digital back door unlocked, allowing a hacker to commandeer the entire fleet.
Promoting “Platform Decay”
The existence of Doublespeed highlights a blatant violation of the social contract—and the Terms of Service—of nearly every major social platform. Meta (Instagram and Facebook), X, LinkedIn, and Reddit all strictly prohibit posting at high frequencies or buying and selling engagement. Yet, enforcement has become a losing battle.
Tech critic Cory Doctorow refers to this phenomenon as “platform decay”—a state where online spaces are infested by bot accounts and low-quality, AI-generated “slop.” Users are becoming increasingly frustrated as the authenticity of the internet dissolves. Ironically, while startups like Doublespeed pump AI-generated noise into the ecosystem, the platforms themselves are cutting costs by using AI-powered moderation to filter it out. It is a snake eating its own tail: machines creating spam for other machines to moderate, while human connection is squeezed out.
Innovation or Exploitation?
The Doublespeed saga serves as a peak indicator of the current tech zeitgeist. It represents a million-dollar gamble on a for-profit company whose entire model relies on monetizing the very thing that makes the internet unusable.
The hack didn’t just expose security flaws; it exposed the ethical vacuum at the heart of the “AI influencer” economy. When venture capital flows into automating deception, forcing algorithms to serve “human” content that is actually generated by a rack of 1,000 phones in a server room, the definition of innovation is stretched to its breaking point. As noted by critics, this may be the ultimate answer to the question: Who said capitalism has to breed innovation? Sometimes, it just breeds more spam.