Hacker News Reader

Comment by 1024core | Hacker News Reader

While Waymo is spending $$$ gathering driving data, Tesla has 100s of 1000s of cars doing it, for free. In terms of sheer data, in this race, Tesla wins. Now whether Tesla can actually use the firehose of data and actually train models that use it productively, remains to be seen. With the departure of Karpathy, I am not so sure.

If Tesla gave all the data to Waymo, Waymo would reach L5 in no time.

View on HN

Waymo was collecting data before Tesla, and switched to simulated training a long time ago because it's more effective.

The self driving AI can gain 100 years experience in just 1 day using simulation: https://blog.waymo.com/2020/04/off-road-but-not-offline--sim...

Tesla uses simulation to but there is no replacement for real world data. The real world is crazy and Tesla can see people driving in real condition from Alaska to Miami Florida.

Amateurs think that "simulation" is all you need. They see realistic environments on their PS4s and XBoxes and think that's enough.

If simulated data were enough, why _hasn't_ Waymo launched in placed like NYC or Boston?

If you think driverless car training in simulations is a playstation game, then you're projecting with your comment. Also, because of weather.

Waymo just went live with full driverless for the public in Phoenix. Tesla can't do full driverless so I think that evidence settles the debate of simulated vs. data collection, unless you have evidence otherwise?

If the goal was to spend 10 billion on establishing a money losing operation in Phoenix then maybe you would have point.

But in fact that isn't actually why 10 billion $ were invested.

Waymo is the first fully driverless vehicles to market, which Tesla has yet to deliver, so the fact you calling it a "losing operation" shows your confirmation bias here. Did you buy a Tesla with this promise, perhaps?

You still didn't answer the question - where's the evidence?

Seems like something I'd expect CNBC to say, not Hacker News.

Hacker News is superficial negative group think. Inverse hacker news is where it's at. Then you can predict the success of startups like Dropbox and Coinbase

I do have some experience in building ML systems, and know enough to know that I don't everything. But I do know the value of real world data in a noisy environment like driving conditions.

Reply on news.ycombinator.com