You may remember the Tesla CEO last year trying to publicly shame a “tester” for being critical.
The CEO went further in the thread to whine that giving early access to software means testers will be very harshly criticized if they dare to criticize.
Such dictated dysfunctional displays of communication (Tesla wants boot-lickers only) are symptomatic of their product failures.
On that note, Tesla recently issued a recall that publicly admitted their fraudulently named FSD is a failure (“anachronistic and just flat wrong“). They argued failing to deliver what they charge for in advance fees… isn’t plain fraud (e.g. advanced fee fraud).
In a similar fashion when their door handles stopped working, Tesla argued they don’t have to fix them after basic warranty because they don’t consider it a defect when their ideas/designs are huge failures that lead to more advanced fees for more failures.
See? You can’t criticize failure because that’s what they’re selling, with no obligation to deliver anything that works. It’s like the Nigerian 419 scam except Tesla victims are far more likely to be trapped and burned to death after being tricked out of their savings.
For example, the criticism above is that Tesla has over-focused on a high profile complicated unprotected left turn instead of more basic safety issues.
It looks like an orchestrated PR move to swing sentiment and hide reality of defects. An instrumented product engineering decision based on expertise in reducing failures likely wouldn’t have poured resources into promoting dangerous left turns.
I’m reminded of the tragic Tesla decision to push FSD with inexpensive least capable equipment and no real upgrade plan (after both Mobileye and NVidia walked away, citing toxic management).
A report from The Washington Post reveals that a number of Tesla engineers were “aghast” at Musk’s insistence on removing radar, widely seen as a cost-cutting measure. Engineers feared that the removal of such important sensors could lead to an uptick in crashes, noting that the cameras couldn’t be relied upon if they were obscured by raindrops or bright sunlight.
The flailing industry-laggard Tesla AI model is completely inverse to common sense, generating huge amounts of extremely low quality data as a big bogus show of “working hard” instead of trying to get signal through more trustworthy high efficiency investments that would make real measured progress.
It’s like when someone dumps tens of thousands of heavy paper pages (mostly blank) on a table and boasts they’re the smartest person in the room working on the hardest problems, instead of a normal person offering a 10 page report showing actual intelligence.
Tesla solving 1+1=2 a hundred million times (lower level or no level) is NOT equivalent to someone doing basic algebra (higher level) a thousand times. Yet you’ll see Tesla often arguing they have the “most” miles as a form of trickery, emphasizing again they are selling failure (e.g. they’re still getting 1+1=5).
Worse, Tesla PR using social media about failing a calculus test repeatedly doesn’t validate any massively high volume low quality low level program. A proud parent boasting their baby soon will do quadratic equations is… nonsense. Focus instead on changing that full diaper maybe?
Tesla failures since 2016 have suggested fraud: critics silenced in favor of human guinea pigs in PR stunts caught up with constant broken promises.
So, as I pointed out before with “beta” version 10, the latest release of software to public roads is putting innocent people in harm’s way. FSD 11, despite all the big claims to focus on the high challenge of automated unprotected left turns, immediately is being demonstrated by “testers” as unsafe.
Start at 13:15 and watch FSD 11 creep into an intersection, blocking a crosswalk to run a red light.
Ladies and gentlemen, Tesla’s best attempt ever still doesn’t see red lights in intersections.
The test driver warns things only get much worse after that, such that by 38:45 he is angry and says “face palm… one blunder after another… really uncomfortable… no point in using FSD”.
One thought on “Tesla FSD 11 a Failure: Hyperfocus on Bad Left Turns Doesn’t Deliver”
I want a car that feels like a WWI biplane in nosedive that I have to fight to correct or die.