Driverless car startup Cruise's no good, terrible year
(www.npr.org)
from L4s@lemmy.world to technology@lemmy.world on 31 Dec 2023 06:00
https://lemmy.world/post/10150177
from L4s@lemmy.world to technology@lemmy.world on 31 Dec 2023 06:00
https://lemmy.world/post/10150177
Driverless car startup Cruise’s no good, terrible year::undefined
threaded - newest
This is the best summary I could come up with:
As 2022 wrapped up, CEO Kyle Vogt took to Twitter to post about the company’s autonomous vehicles rolling onto the streets of San Francisco, Austin and Phoenix.
Local reports showed footage of confused vehicles clogging a residential cul-de-sac, driving into wet cement at a construction site and regularly running red lights.
They tallied nearly 75 incidents where self-driving cars got in the way of rescue operations, including driving through yellow emergency tape, blocking firehouse driveways, running over fire hoses and refusing to move for first responders.
“Our folks cannot be paying attention to an autonomous vehicle when we’ve got ladders to throw,” San Francisco Fire Chief Jeanine Nicholson said in an August hearing.
Based on police reports and initial video footage from Cruise, the woman was first struck by a hit-and-run human driver whose vehicle threw her into the path of the driverless car.
The new video footage showed the Cruise car striking the pedestrian, running her over, and then dragging her an additional 20 feet at 7 miles per hour as it pulls to the curb and stops on top of her.
The original article contains 1,352 words, the summary contains 183 words. Saved 86%. I’m a bot and I’m open source!
I’m not sure why I feel the need to preface this comment, but, I work in software, I get how hard a problem autonomous driving is, how genuine safety improvements over human drivers are highly valuable, and how perfect need not be the enemy of good.
However, the level of sheer blind optimism from the AV crowd is the same as the AI “leaders” and the crypto bros before them. How their statements are not straight up fraud is beyond me.
The reality of them needing to have a remote team of drivers intervening every 2 to 5 miles of driving, within an urban setting very much designed for cars, is so far away from the picture they paint.
No wonder the tech industry has a dog shit reputation.
People have been saying since the 1950 that this time they really cracked the code of AI and soon it will (enter lofty claim here). I mean some serious steps have been made but in the end it never lived up to the promises made. I’ve got no reason to believe that this time will be different.
I mean, I think it’s a little different in that there’s tangible AI products that millions of people are already using?
I have my own doubts about how the current architectures scale towards “general intelligence”, but seems like a very real breakthrough that is already producing at a significant level.
I’d describe those products as being in the “heh, that’s neat” phase. At least, the good ones anyway. After a few hours you start the see the issues - like images with very weird fingers - and then you the illusion is broken.
For example, in work I’ve been playing around a lot with various LLMs to generate marketing copy for physical products. I’m being vague here for reasons I’m sure are obvious.
Across the first few hours it felt very impressive, it would pretty much instantly churn out descriptions in a variety of styles, even when the product information we provided was low quality.
The problem though, was that the copy it wrote was actually - if you sat down and read it - shit.
Now, the vast majority of marketing copy is also shit. The good stuff is excellent, rare as hens teeth, and incredibly expensive, but your generic boilerplate crap you see all over the Internet, that stuff, it could replicate perfectly. Even the lies.
If you want sales speak waffle, it’s 11/10 every time.
So we’re currently in a bit of a bind. Do we release a tool that creates boilerplate shit more easily, and will turn even the most inconsistent (and straight up conflicting or impossible) data in to fancy sounding text, and jump on the AI-powered bandwagon, or do we not, because what it creates isn’t actually any good?
Imo spitting out crap in a second or two isn’t a valuable improvement on where we were a few years ago, even if it is pretty neet.
I agree with what you’re saying about your line of work. I code for a living, and Copilot is genuinely useful all day long. I use it now and again to generate a script from scratch, but most of the time it functions as either an incredible autocomplete of whatever I am coding, or it converts a chunk of code from one format to another, with just a description of what I’m trying to do, instead of me having to write a complex regular expression or do it by hand. So I’d say in my line of work it’s gotten past the “neat” phase.
But it still ultimately requires you in that chair to correct issues.
I don’t doubt for a second that time is saved, especially in the boilerplate parts of writing code, but it’s not going to remove you from that chair.
From my experience, Copilot is helpful, but we’re talking a few steps above templates, Clippy 3.0, if you will. The bit that blew my mind a bit was realising just how much code I write is the essentially the same. “Automating” that has been a great help, but - like with the marketing stuff - it’s not at a point where it can do it alone.
It’s the same with driving, etc.
There are lots of awesome things that help me with my job but still require me to be there. But you were saying the help it gives with the job is more limited than it seems to be at first. And I’m just saying in my line of work it’s actually a huge help.
It’s not like Clippy or templates. I have to spend time setting up templates and following a specific structure and syntax, remembering to use them. With Copilot I turned it on one day and it was instantly helping me with whatever I was working on, and continues to do so no matter what comes up.
I wasn’t making any assertion that it could do my job without me, but it seems far more useful in what I do than what you had described.
The Cruise propaganda video that Derek from Veritasium put out on YouTube is what finally got me to unsubscribe/block his channel. Cruise was using shady tactics way before these safety incidents.
Relevant Vid: youtu.be/CM0aohBfUTc
I was with it until the SurfShark promo bit, which, painfully ironically, has almost-lies and ‘yes but actually no’ bits all throughout it. Dude shoots himself in the foot.
I had no idea. I use SponsorBlock to avoid seeing sponsored segments like that one. Thanks for the heads-up.
I’m on my phone using Grayjay, and it doesn’t (yet, hopefully) offer SB integration; I’d usually just fast forward, but I was like ‘huh, wonder why he chose a, ehem… popular sponsor for a video like this’ and sat though it, thinking he’d actually dissect the advertising-speak, but nope. Bah :(