Hacker News Reader

Comment by programmarchy | Hacker News Reader

Musk's Tweet about FSD is pretty much spot on. Just being electric isn't enough of a market differentiator anymore. Too many copycats. Sacrificing a new model to focus on FSD is a great move. Many stockholders probably won't be able to stomach it, though.

> Not quite betting the company, but going balls to the wall for autonomy is a blindingly obvious move.

> Everything else is like variations on a horse carriage.

https://twitter.com/elonmusk/status/1780376546148327690

View on HN

> Sacrificing a new model to focus on FSD is a great move.

Musk has been saying level 5 fsd is just around the corner for well over a decade, demonstrating a large and persistent disconnect between expectations and reality. That makes the reward/risk ratio of sacrificing a model to sink more into fsd rather questionable. The outcome isn't just not guaranteed, it isn't even known to be solvable with today's technology.

People will buy nice electric cars that aren't riddled with quality issues. Investing in QA (where Tesla has next to none) is a bog-standard, practically-guaranteed route to improving products. But that doesn't excite this "brilliant leader" so it's off the table.

This sacrifice isn't a "great move," it's another mark in Musk's record of bad gambles.

> People will buy nice electric cars that aren't riddled with quality issues.

Anecdata: my father bought a VW ID.4 1 year after I got my M3 (2019). The MY wasn't available yet. VW ID.4 has been at the garage 3-4 times/year for various (mostly software) issues. Same with my mother's Peugeot e-208. My M3 got the latest software update the next day after an imperfect firmware update made autopilot sometimes unavailable, all done with the technicians via the Tesla app.

You can choose between amazing efficiency and software, and perhaps a 1mm-off panel on 1 out 1000 vehicles, or perfect esthetics but horrible software.

Teslas are still not perfect, and a stronger QA would definitely make for a better image, I agree. Still, Tesla has made a lot of progress, you don't find as many people complaining about QC/panel issues anymore on reddit. But competitors still have a ton of catching up to do.

The eternal optimism of Tesla fans continues to astound me. The cybertruck has just been recalled because the soap they used to install accelerator pedals made them slip off, and you're telling me people are complaining less about panel issues?

Other manufacturers have recalls too, for equally dangerous conditions, but only Tesla ever makes the news (I'm not defending Tesla, just proving a point). CT has been out for ~4 months, 4k units delivered, platform engineered from scratch, new low voltage system, steer by wire, ...... inevitably something will slip.

Compare a brand-new everything to minor updates to existing cars:

- https://www.cbsnews.com/news/toyota-lexus-recall-2020-2022-f...

- https://www.cbsnews.com/news/toyota-recall-50000-corolla-rav...

- https://www.cbsnews.com/news/toyota-recall-rav4-suv-2023/

- https://www.cbsnews.com/news/toyota-recall-tundra-fire-risk/

And this is just Toyota, which as far as I can gather is considered reliable and trustworthy.

> only Tesla ever makes the news

This bald-faced lie is revealed by the very news you share further down your comment.

> (I'm not defending Tesla, just proving a point)

Your argument takes the form of a whataboutism, which does not actually prove anything.

> Musk has been saying level 5 fsd is just around the corner for well over a decade, demonstrating a large and persistent disconnect between expectations and reality.

You literally have Musk's Director of Engineering telling/"reassuring" the State of California that Musk's statements on FSD are "visionary, and not reflective of engineering reality".

Brilliant

When were they not focusing on FSD? It was promised like 5 years ago.

This is just a rationalization as to why the car company isn't growing, despite people claiming they'd own the entire market.

I do wonder if he's doomed FSD/robotaxi by being stubborn about lidar. The self-driving problem is unlike most other ML problems due to its mission critical nature. Having 2 modalities that can be cross-checked will reduce edge case hallucations that cause crashes. Like, instead of probability p that you don't see the truck in the middle of the road, you reduce it to nearly p^2. This non-linear reduction in faults that comes with cross-checking modalities is not intuitive. I never bought Musk's justification: "Lidar is not how humans do it". Why would we intentionally handicap ourselves to the tools available to biological evolution?

Seeing how hard it's been for Cruise and Waymo, who use lidar, and only operate in a small area, I am not expecting a functioning Tesla robotaxi that regulators will allow on the road anytime soon.

> Sacrificing a new model to focus on FSD is a great move.

They are actually not doing that. TSLA said on the earnings call they are accelerating the completion of the cheaper Model 2 platform. That's what driving up the stock price AH.

https://finance.yahoo.com/news/tesla-earnings-q1-175358835.h...

Focusing on FSD is car company suicide. It's a focus on selling cars to people who don't like cars and would rather do without one entirely. You can sell some cars like this but ultimately it's a dead end, you soon sell one to everybody who wants one, nobody else does, and the market for your cars crashes.

This has been covered in the past. Some people may own their car but the robotaxi model folds into transportation infrastructure, like Uber. Cars sitting in parking lots and driveways 90% of the time is incredibly stupid and wasteful. Not to mention commercial transport. The market is enormous.

Why anyone would take what he says about FSD seriously, and with any shred of hope attached, is beyond me at this point. What he says here is logical, but he has so consistently failed to deliver I can’t fathom continuing to have faith in what he says about FSD.

It literally went from being a solved problem in 2015 - "I view it as a solved problem. We know exactly what we need to do and we will be there in a few years” - to somehow getting un-solved along the way. In 2022, "Our focus now is just on working on solving this problem".

Classic 90/90 rule. Not surprising to anyone who has done innovative engineering. The last stretch is always the hardest.

"Innovative engineering?" I think you mean "fundamental research." The classic ∞/100 rule?

I think most realistic beliefs are that FSD is a minimum of a decade away (personally, my belief is more than, and when it does come to life, it won't be Tesla)...

Anyone who says "it's a solved problem, we just need to implement it" is deluding themselves, and as soon as you poke... it's qualifier, after qualifier, after qualifier. Tesla isn't in "the last stretch". Navigating around Pittsburgh at night in winter blizzard conditions, with poor roads and signage, whiteout over what street markings there are, and poor visibility?

That's the last stretch, not "it handles roundabouts far better than it did in FSD version x-1!".

If it’s not surprising to anyone who has done innovative engineering, how did Elon miss it?

Stock is up almost 7% after hours. Someone is believing the robotaxi/autonomous driving narrative.

FSD just killed a guy today.

Worth mentioning: I don't believe they have confirmed it was FSD yet. It just says Autopilot which can be the basic one which just keeps you in a lane.

In any case it's frightening that any version of it it wasn't capable of detecting a motorcycle in front of the car and stopping.

It probably was FSD -- FSD is available to pretty much everybody in April, so it seems unlikely he was using Autopilot.

But what seems likely to me is that the driver had his foot on the accelerator. FSD is relatively timid. For example, it will slam on the brakes if it sees a pedestrian that it thinks might cross in front of the car. The driver can press on the accelerator to prevent this. I've gotten used to just keeping my foot on the accelerator for one particular road where FSD is particularly prone to slowing down for no good reason.

“Copycats” is a bit rich.

So we get neither a cheap Model 2 nor actual FSD.

Just like how Trump creates some new sensation when he's in a tight spot, Musk creates some new vaporware. In this cases he's even resorting to recycling existing vaporware.

"Full Autonomy in 9 months! No this time I mean it!"

Reply on news.ycombinator.com