‘Completely Drove Straight Past Us’: Man Hails Waymo Late-Night. Then It Changes Its Pickup Location 4 Times
“So, we changed the pickup location three times. Actually, we didn't. The Waymo did."
If an autonomous car changes its pickup location four times before you ever get in, is it still “reliable”? That’s the question one TikTok user accidentally raised after filming their first late-night Waymo ride home from an Addison Rae concert.
TikToker Mason (@masonncarmona) and friends appeared excited about their first self-driving rideshare experience after leaving a concert. But their enthusiasm waned a bit in the clip, which has been viewed more than 500 times, when the robotaxi unexpectedly changed their pickup location again and again.
“So, we changed the pickup location three times. Actually, we didn't. The Waymo did,” Mason said before one last unexplained adjustment from the Waymo. “And so now we have to go all the way over here around the block, because it changed it to the block that's like across from us.”
Now Trending
For readers well versed in electric vehicles and autonomy, this moment may seem small, even amusing. Yet it touches on a key question: When the “driver” is software and sensors rather than a human, how much of the service is about safety, and how much is about predictability and user experience?
The incident doesn’t involve any obvious safety issues; the vehicle arrived and completed the ride. But it highlights that reliability in an autonomous-ride context is about meeting user expectations, minimizing wait time, and behaving in a predictable manner.
Waymo’s Safety Credentials
When it comes to crash statistics and injury prevention, Waymo’s numbers are impressive. The company reports that its “Waymo Driver” system has achieved reductions of 91% in serious-injury-or-worse crashes, 80% fewer injury-causing crashes, and 79% fewer airbag-deployment crashes compared to human drivers in its operating cities. A study in collaboration with reinsurer Swiss Re, which covered 25.3 million fully autonomous miles, found an 88% reduction in property damage claims and a 92% reduction in bodily injury claims compared with human-driver benchmarks.
FROM THE TRENDING NEWS DESK
Viral bits from across the social media landscape
Our team of experts tracks what's trending so you don't have to—from viral videos to online debates that have everyone talking.
Another more recent peer-reviewed analysis covering 56.7 million rider-only miles found statistically significant crash-rate reductions across 11 crash-type categories, including 96% fewer injury-involving intersection crashes. These data strongly suggest that on the metrics of safety—crash reduction, injuries, vulnerable-road-user protection—Waymo’s driverless system is demonstrably improving over human driving in the zones where it operates.
For a consumer riding a robotaxi, reliability also means answering other questions: Did the car show up when and where I expected? Was the pickup location clearly communicated? Did the system adapt smoothly to real-world conditions, including concert egress, heavy pedestrian traffic, and changing road conditions? The reported pickup-zone changes in Mason’s ride draw attention to these operational dimensions.
Even if safety is the first line of rational trust, user-experience friction like walking to a shifted pickup point or waiting longer than expected can erode confidence. And for a system that aims to compete with human-driven rideshares, these small glitches matter.
Mapping, Environment, And Real-World Complexity
Robotaxis operate under constraints. Most commercial autonomous rides today are geofenced within specific urban zones, with known mapping, controlled speeds, and predefined operational parameters. When you throw in “late night outside a concert,” you introduce variables: shifting pedestrian flows, unfamiliar drop zones, changing lighting, complex surroundings.
In the recorded ride, the changing pickup location may reflect behind-the-scenes repositioning logic, vehicle movement to optimize pickup, algorithmic corrections or mapping nuances in which the system selected a safer pickup point based on sensor constraints. While we don’t have the internal log for Mason’s ride, the episode underlines how real-world rides are still evolving beyond laboratory conditions.
The story is instructive on several fronts. First, the EV/autonomous pairing is not just about drivetrain and battery; it’s a system of systems: mapping, sensor redundancy, user-app experience, operations, logistics. Waymo’s strong safety numbers give credibility to the EV+autonomy future, but the pickup-logic glitch shows that operational refinement remains critical.
Second, deployment context matters. Waymo likely optimized its safety and operations in its initial deployment zones, such as Phoenix and San Francisco, over tens of millions of miles. But every new use case can surface fresh challenges, whether mapping gaps, crowds or unexpected pickup shifts. As autonomy scales, these edge cases become growth pain points.
Third, communicating reliability to users is as important as delivering it. A user might say, “Yes, the ride was safe,” but also, “No, I had to walk around the block.” For an autonomous rideshare service to gain widespread trust, the seamless experience must follow (or approach) that of human-driven alternatives.
In the narrow sense of safety outcomes—including negative ones such as collisions, injuries, and liability claims—Waymo’s autonomous fleet shows meaningful, statistically significant gains compared to human drivers.
But in the broader sense of “booking a car at midnight, from a crowded event, and trusting it to show up exactly where I’m standing without a hitch,” the TikToker’s video reminds us there’s still room for improvement.
InsideEVs reached out to Mason via email and direct message. We’ll be sure to update this if they respond.
RECOMMENDED FOR YOU
Woman Buys Cadillac EV. Then They Offer Her A Deal On EVgo
The Skoda Epiq Is Europe’s Big Small Electric SUV
"The Motor Is Supposed To Last 1 Million Miles": Man Hits 360,000 Miles On His Model 3. Is That The Norm?
ChargePoint Wants To Fix The Condo EV Charging Problem With 2,500 New Ports
"Dealerships Are Shady": Man Says Texas Toyota Dealership Agreed To Sell Highlander Hybrid For $26,900, Then Jacked Up The Price
Costco’s New Discount Makes America’s Cheapest EV Even Cheaper
Can Waze Still Track The Police?