Профиль

OP didn't say and it is not obvious how that robot got so far off. 2nd Street in that direction has only one car lane and Bryant is one-way w a right turn that could only get there if made from the wrong/not-rightmost lane. fwiw, seeing rando sensor rigs on poorly marked cars is old/common in SF.

still a favorite. Takes much less time and money to train a normal person or even thousands of them to handle this situation than a robot. bsky.app/profile/anic...

Not even the first WWW. Note there used to be a WWW video ~monthly, then ~weekly, now ~daily. Long arc of WWW anecdata bends toward ~constantly/continually. bsky.app/profile/anic...

Yet another (133+) Wrong Way Waymo: 2nd at Bryant, San Francisco Not the first, second, third, or fourth Waymo robot to screw up and/or drive illegally at/around this intersection. Thanks for coming to my evergreen TED talk.

fine print, one Product Communications Manager at Waymo tweet at a time, bc this may not be a question they trust Gemini to answer for itself.

Most in-depth article I've seen of what role Waymo's robotics hw may have had in the fire when their robot was rear ended in Atlanta. I'm not aware of any Federal or State regulations or tests of the crash or fire worthiness of extensive ADS vehicle modifications. theautowire.com/2026/07/29/w...

I suspect Waymo's brilliant UI team's FAFO/feature wishlist is running out while they wait for Waymo's most brilliant AI/sw team to fix bugs they haven't in >17 years and Waymo's brilliant ops team to get their act together in just about every way imaginable just about everywhere not a test track.

Waymo put Google's AI "assistant" Gemini into their 6th gen robot PAX UI to offer "new ways to interact with...everyday knowledge," in the words of their CPO Saswat Panigrahi. In case you were wondering about Plan B to make up for their robot's defective "everyday knowledge." Plan A being telops.

"In Sacramento, City Councilmember Phil Pluckebaum said Waymo's rollout is likely just a month away. "They've got this down, they've done this before, they know these roads," Pluckebaum" The AV/ADS education of Phil Pluckebaum begins and not well. Waymo needs CPUC's ok to charge for rides there.

NHTSA's ADS crash reporting change in June 2025 appears to have eliminated ~half give or take of the non-injury crashes, at least for Waymo. Note: ~60 crashes per ~800kVMT went to ~30 and projects to ~60 crashes in ~1400kVMT. Good to have friends w fingers on the scales of your most PRed stats.

There are rules (science & eng/ops best practices) to making hard realtime systems w remote operators work that trace back >100 years to scaling the telephone and updated all the way to interplanetary systems. A consumer-grade public road lifesafety system is among the worst possible applications.

This is and has all along been THE big AV story none of these companies want told and certainly not regulated w minimum performance standards set by anyone but themselves. Concerns go all the way to the top of fleet ops, eg Waymo's has all but caused their worst outages. bsky.app/profile/anic...

The telops lies/fantasy every AV co (Waymo, Zoox, Tesla, Cruise, etc) tells: 1) robot always knows when it needs help 2) telop is always available 3) telecomm is always adequate 4) telop is always effective (make better/safer) Plenty of incidents and videos of each of these failing dangerously.

Fired Telsa "robotaxi" telop mgr alleges overwork. China requires =<3:1 robot to telop limit. Any ratio >1:1 guarantees times of overwork endangering the public, as Waymo has repeatedly demonstrated. The entire AV industry rests on this ill-conceived hack. www.automotiveworld.com/news/new-law...

Waymo is "welcoming riders on freeways again after...software updates...first in Phoenix. San Francisco, Los Angeles, Austin, and Miami will follow" Sorry Atlanta, Nashville, Orlando, Dallas, Houston, & of course San Antonio. Took Waymo >month to trust/test this update. archive.ph/kXO3W#select...

If the robot stopped in a through lane for no good reason, it could be defective. Waymo's NHTSA crash report should provide a reason, expected public ~mid-September. Other robot companies have had recalls for stopping for no good reason leading to being rear-ended, eg Amazon Zoox first recall.

put a laminated copy in every robotaxi by the call button

The truck wasn't speeding and the Waymo robot did not have a stop sign. Waymo has not explained why their robot decided to make a full stop before making the turn.

"what others are hiding" is that there are no others, effectively. FSD is ADAS not ADS. NHTSA requires reporting crashes when automation was engaged within 30 seconds of impact. Only Waymo & GM Cruise have operated without a safety driver for >million VMT in US. It didn't end well for Cruise.

Waymo reports crash severity to NHTSA. NHTSA changed non-injury crash reporting requirements in mid-June 2025, effectively lowering the count, so there is a discontinuity there. California non-injury crashes appear to increase ~linearly w ridehail trips. Injury crashes have been ~9% of crashes.