Tesla's Dangerous Doors (electronic door latches) - Can Rivian please not follow this trend?

[VandalSibs]

You shouldn’t have to go digging for that information online when you’re trying to decide if you want to buy the car or not.

That kind of thing is part of my research when buying a car, but everyone has their own method of shopping.

Sponsored

 

[JohnB R1T]

Auto Lock (when walking away) can be aggravating and has its own "dangers"...If I'm packing the truck and away from "home", I generally pop the driver's side gear tunnel open slightly. The truck won't lock itself. Confession: I usually forget that I've done this and then have to get out and close it when the truck won't go anywhere (because it's open).

The "danger" is that my wife's Audi doesn't have this auto lock feature
(you also have to push the "start button" to turn off the engine, otherwise it stays on for an hour)
(since it's a hybrid the engine isn't "running" or making noise, it's just "on").

I've gotten out after failing to turn off the engine, then walked away after failing to lock the door --only to have my wife chastise me when the car tells her that it's unlocked via the app.

 

[srkz]

Geez. If there is a fire the responding fire crew isn't going to bother retrying door handles—no matter what type you have. If they can't get the door open right away, they're ripping it open like a tin can with this. Don't fall for sensationalist journalism.

I have no doubt the responding crew will be successful in retrieving the bodies. The goal is for everyone to get out before burning to death.

 

[slillie]

You shouldn’t have to go digging for that information online when you’re trying to decide if you want to buy the car or not.

Actually, before purchasing a $100,000.00 vehicle, I do a lot of research: "The Google," ChatGPT, Grok, any AI App. It's all there and easy to find.

 

[Davethadog]

Actually, before purchasing a $100,000.00 vehicle, I do a lot of research: "The Google," ChatGPT, Grok, any AI App. It's all there and easy to find.

Relying on historically inaccurate wastes of electricity is a stupid strategy when you want to know how to quickly extract yourself or a child from a malfunctioning vehicle. I visited the space too and they didn’t even know. It’s not unreasonable to expect an employee who explains the car to people for money to be able to articulate how to do it. Rivian shipped the earliest gen 2 vehicles with explicit info on how to but pulled it before delivery - that’s fucking lame.

Here’s what happens when you ask ChatGPT how. Ask 3 barely not retarded LLMs, get 3 different equally as inaccurate answers.

 

[Dave Cundiff]

In the last two days, I've found Google's AI response to web searches to be factually wrong, about facts that were already known to me and/or clearly presented in the web results below.

***

From what I understand, large language models (LLMs) make stuff up for at least three reasons:

(1) They are programmed to please the humans with whom they interact. They will make up, or distort, their output to tell you what they calculate is most likely they want to hear.

(2) They scraped others' output to develop their own. When "garbage in" predominates, so will "garbage out." Unlike educated humans, LLMs lack reliable ways to evaluate the quality of what they have read.

(3) Feedback to LLMs is always at risk of being dominated by "how plausible does it sound?" Plausibility is easy and cheap to evaluate. Accuracy is relatively difficult and expensive to evaluate.

***

If you trust LLM/AI output for anything related to safety or finances -- yours or anyone else's -- you are at risk of being harmed or of harming others.

Very best wishes!

Sponsored