Tech

Tesla hits Musk’s threshold for ‘safe unsupervised’ driving

Tesla’s 10-billion-mile milestone for Full Self-Driving (Supervised) triggers Elon Musk’s long-promised “safe unsupervised” threshold—yet the cars remain stubbornly Level 2, still demanding constant human oversight. The gap between marketing benchmarks and regulatory reality widens as Tesla’s fleet data outpaces its actual autonomy stack. AI-assisted, human-reviewed.

Tesla's fleet of vehicles using the company's Full Self-Driving (Supervised) system has driven over 10 billion miles, according to the company's updated safety page. This crosses the threshold Elon Musk set in January for "safe unsupervised" driving. However, FSD remains a Level 2 system requiring constant human oversight, and Tesla owners have not gained access to unsupervised driving.

The threshold

In January, Musk stated on X that "roughly 10 billion miles of training data is needed to achieve safe unsupervised self-driving." The implication was that once reached, Tesla would enable unsupervised driving for customers. The company has now hit that number, but the cars remain Level 2 — drivers must keep hands on the wheel and be ready to take over at any moment.

Liability questions

Tesla's terms of service place liability on the owner for crashes involving FSD, characterizing it as a Level 2 supervised system. If FSD were unsupervised, who would assume responsibility? Waymo, which owns its tech and fleet, assumes liability for its vehicles. Tesla has faced hundreds of crashes involving its partially autonomous features and dozens of fatalities, but has largely avoided liability through settlements or dismissed lawsuits. A federal jury in Florida last year found Tesla partly liable for a 2019 crash involving Autopilot and ordered the company to pay $243 million; Tesla appealed, but a judge rejected that effort.

Safety claims and methodology

Tesla claims its FSD-equipped vehicles drive 5.5 million miles on average before a major collision, versus 660,000 miles for the average US driver. Experts have questioned this methodology. Studies indicate Tesla's safety reports fail to account for basic traffic statistics — for example, crashes are more common on city roads and undivided roads than on highways, where Autopilot is most often used. Some researchers believe Tesla may be miscounting crashes to make Autopilot and FSD appear safer than they are.

Robotaxi fleet progress

Tesla is ramping up its unsupervised robotaxi fleet. After launching in Dallas and Houston with two vehicles each, Dallas now has five unsupervised robotaxis and Houston has six, according to the Robotaxi Tracker. Austin, where Tesla first launched its robotaxi service, now has 29 supervised vehicles (with employees in the front passenger seat) and 22 unsupervised ones.

What's next

In an earnings call last month, Musk said unsupervised driving would come when "it is legal to do so." Asked specifically about unsupervised FSD in customer cars, he predicted it would arrive in the fourth quarter of the year. Whether this is another threshold or another moving goalpost remains to be seen.

The practical takeaway: Tesla's

Similar Articles

More articles like this

Tech 1 min

Elon Musk sent ominous texts to Greg Brockman, Sam Altman after asking for a settlement, OpenAI claims

Elon Musk's alleged ultimatum to OpenAI's leadership, involving a veiled threat of public backlash, has surfaced in a contentious dispute over a proposed settlement. The texts, reportedly sent to President Greg Brockman and CEO Sam Altman, hinted at a potential smear campaign targeting Musk and Altman. The exact nature of the proposed settlement remains unclear, but Musk's ominous warning suggests a high-stakes negotiation. AI-assisted, human-reviewed.

Tech 1 min

The best Star Wars Day deals

Star Wars Day’s retail blitz turns nostalgia into a loss-leader arms race, with Amazon, Walmart, and Best Buy slashing prices on 4K box sets, Lego Ultimate Collector Series kits (up to 30% off), and even branded Band-Aids—weaponizing fandom to clear inventory ahead of the franchise’s fallow period between *The Mandalorian & Grogu* and *Fate of the Old Republic*. AI-assisted, human-reviewed.

Tech 2 min

Anthropic and OpenAI are both launching joint ventures for enterprise AI services

Two leading AI research firms, Anthropic and OpenAI, are expanding their reach into the enterprise market through strategic partnerships with asset managers, effectively bundling their large language models with financial services to appeal to high-stakes clients. These joint ventures will integrate AI-driven decision support tools with traditional investment products, targeting institutional investors and corporate treasuries. The move signals a growing convergence of AI and finance. AI-assisted, human-reviewed.

Tech 2 min

Getting Digital Fairness Right: EFF's Recommendations for the EU's Digital Fairness Act

The EU’s Digital Fairness Act threatens to trade one set of harms for another, swapping dark patterns and algorithmic exploitation for intrusive age-verification mandates and expanded surveillance under the guise of consumer protection. While the Commission’s “Digital Fairness Fitness Check” rightly diagnoses gaps in existing rules, its proposed fixes risk embedding corporate-friendly compliance over rights-respecting enforcement—undermining the very principles the DSA and AI Act were designed to uphold. AI-assisted, human-reviewed.

Tech 1 min

As X shuts down Communities, Acorn debuts an alternative that puts creators in control

As centralized social platforms falter, decentralized alternatives like Acorn's community-as-a-service platform are gaining traction, empowering organizations to build and manage their own online forums using blockchain-based moderation tools, custom feeds, and granular analytics, potentially upending the traditional social media paradigm. Acorn's solution leverages a modular architecture and open standards to give creators control over their online spaces. This shift towards decentralized social infrastructure may redefine the future of online community building. AI-assisted, human-reviewed.

Tech 1 min

Brussels reissues its Huawei warning, six years on, and prepares to make it stick

Brussels is poised to codify its six-year-old Huawei and ZTE ban into enforceable law, transforming a voluntary 5G security warning into a continent-wide legal firewall. The move, backed by threat intelligence on supply-chain backdoors, would bar the vendors from core and radio-access networks—triggering Beijing’s vow of “proportional countermeasures” against European telecoms and exporters. AI-assisted, human-reviewed.