Hacker News Reader

Comment by duxup | Hacker News Reader

I can’t tell what this company does.

Does anyone know?

My guess would be PR and myth-building for Elon, et al

I feel like the myth-building stuff really fell apart for Elon after he exposed himself with the Twitter stuff. For a good half-decade in the 2010's he was real-life Tony Stark. Rockets, self-driving cars, clean energy, super unrealistic goals that somehow he kept delivering on despite everyone saying it's impossible, and then then since like 2018 he's just been on a downward spiral. So sad.

I feel the same, but PR works, and so people can change their public persona successfully.

I'm curious where this downward spiral is happening. Sure, Twitter doesn't look to be doing so well. Is one data point sufficient for this spiral?

When I was first reading about Musk, 12 years ago, I was extremely impressed, to the point of being ready to invest in Tesla and SolarCity. However, when Elon entered the AI space about 7 years ago it became clear to me that: (i) he is entering a field that he doesn't know, while talking about it as if he knew it, (ii) marketing and hype matter to him more than science.

> marketing and hype matter to him more than science.

You're assuming he's not exploiting the marketing and hype so he can fund the science that's needed.

I didn't make that assumption. I'm just saying that overpromising and underdelivering, due to a lack of a broad understanding of AI research, can backfire, create distrust, and in the end slow science down, e.g., lead to another AI winter...

Full Self Driving?

Anyone thinking that FSD is evidence of a "downward spiral" isn't paying attention. There's plenty of objective evidence about the state of FSD and while I'm sure there can be broad disagreement about how far from "done" it is, there's no doubt it's trending in the right direction.

Is that why they've stopped publicly reporting disengagement data and stopped seeking certification in California?

Sure, you can guess at motives behind two among countless thousands of business decisions. Or you can observe progress of the FSD Beta software released to customers.

I've had access to FSD Beta since like the very first version. It hasn't made any significant progress to be honest. It still does the dumbest most unsafe things at the absolute worst time. In super ideal conditions it may be ok, but for the most part, I don't even try to use it anymore.

I'm not saying you're wrong, but your assertion is wildly divergent from the experience of many dozens of people who have documented their experiences in full on YouTube. It would be interesting to hear why you think your opinion is correct and theirs isn't. I'd be genuinely appreciative of any thoughts you have.

I'm not saying anyones opinion is incorrect. FSD may work better in some places and worse in others. But I live in Las Vegas. We have wide roads, both highways and regular roads well maintained, great weather, and other elements that would make you think FSD would work great.

In my experience, it does lane keeping great and going in a straight line, slowing down for traffic. But that's about it. Making turns, the overall jank is unbearable in person (it doesn't come off as bad on video, but as a driver or passenger, getting constantly tossed around is not fun), and also just the unpredictability of the system makes it unusable for anything more than a parlor trick.

You're free to have your own opinion or listen to whoever you want, but I can tell you my experience has been consistently terrible and hasn't improved by any meaningful amount over the last 18 months.

Please tell me how I can observe that progress when they have stopped doing any kind of objective public reporting. What are the metrics you deem relevant to track, since not reporting safety data is just a random business decision in your mind?

A count of all-cause disengagements is not "safety data", nor is it "objective public reporting". Much like LoC it can be a signal of progress early on but these days it's ambiguous noise given that many disengagements occur because of reasons other than safety, such as spotting a good place to park moments before reaching a destination. But even where a disengagement is safety related, the data can be confounded — especially given that there has been a rather dramatic uptick in assertiveness in recent releases, revealing further edge cases masked by prior conservativeness.

It's funny, because it reminds me of provisional drivers in my home city of Sydney, Australia. Drivers with less than 1 year of road experience must show a red [P] plate on their vehicle. Drivers with between 1 and 3 years of road experience must show a green [P] plate. This is interesting to me because I consistently find that red [P] drivers are mostly conservative and green [P] drivers are mostly overconfident to the point of dangerous.

Some might say we shouldn't allow people with between 1 and 3 years of road experience to drive on our roads.

Yes disengagements can be noise, but there are significant differences between manufacturers, and Tesla's numbers were not competitive, and then they stopped reporting them, leaving us to draw our own conclusions. It is not a perfect metric, but it is a metric. You never did tell me what metric you are using to judge Tesla's progress. Since you clearly believe they are making progress, maybe you could tell me the reason why you think that.

It will be vaporware, like 90% of what Elon offers.

I remember when half the internet was convinced that the Model 3 was vaporware.

https://news.ycombinator.com/item?id=13216427

Ok, what about: Solar Roofs (fake roofs at demo https://gotmusked.com/content/solarcity-scam-fake-tiles/)

HyperLoop (Not a new idea, has existed for at least 50-60 years, earlier if you look at the atmospheric train)

Tesla Semi (tell us how much they can actually haul, elon)

Cyber Truck (Delayed, and not really very truck-like https://twitter.com/whatsinside/status/1628569254659518464)

Tesla Roadster

Working AI robot at AI day 2

He said they had full self driving back in 2018.

Spaceships to Mars by 2022 - https://www.nytimes.com/2023/04/20/science/elon-musk-starshi...

> "As the name and design have changed, so have Mr. Musk’s overly optimistic predictions for when his spaceship would get to Mars. At Guadalajara, he said the first flight of the Interplanetary Transport System to Mars, carrying cargo but not people, would take off in 2022 and that the first flight with people could launch in 2024."

Most of his promises have no reality backing them

There is exactly 1 opinion stating that model 3 is vaporvawe, on a thread with 215 answers...

The goal of xAI is to understand the true nature of the universe.

They found a way to make philosophy profitable?

Reply on news.ycombinator.com