Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Tesla, software and disruption

ben-evans.com

Post reply on HN
Print view View original

21–30 of 127 posts

Re: Tesla, software and disruption

#21

Post by pbreit »

Animats wrote: post #17

What "massive amount of data" from Tesla? They don't send all the video back to HQ. Tesla owners have observed large uploads, but apparently they are software crash dumps.

It's probably a representation of the video data that has value to providing autonomy. Plus all the path & geo data. With a few hundred thousand cars collecting it, I'd say that's "massive".

Re: Tesla, software and disruption

#22

Post by bri3d »

mrfusion wrote: post #6

If batteries are getting so cheap why isn’t anyone making battery powered: chainsaws, full size vacuum cleaners, jet skis, ride on mowers? All of those would be revolutionary.

Ride-on mower companies are entering the electric space - Ryobi and Troy-Bilt have consumer-level options, Cub Cadet have a prosumer/light pro model (RZT S Zero), and Mean Green make a full line of commercial grade electric mowers.

Electric chainsaws have been a thing for years and have moved from the trimmer size space to the prosumer/light pro space (see the Husqvarna 536Li XP).

Hoover make a full sized appearance canister/upright hybrid vacuum (Air Cordless Lift). In this space, though, I think the prevalence of smaller stick-style cordless vacuums is because the technology enables them rather than because the technology can't be scaled up. The Dyson V6/V7/V8/V10 models aren't great in terms of quality IMO but are very convenient and popular and the V10 is more than sufficient for home vacuuming tasks.

Re: Tesla, software and disruption

#23

Post by bri3d »

Shivetya wrote: post #20
mrfusion wrote: post #6

If batteries are getting so cheap why isn’t anyone making battery powered: chainsaws, full size vacuum cleaners, jet skis, ride on mowers? All of those would be revolutionary.

batteries are getting cheaper but good long term reliable batteries will still set you back a bit. I have had an electric mower, edger, and blower. I only kept the last item. The first just didn't last well enough even with a spare battery and it was already degrading before the end of the season. The edger just didn't cut it, sorry about the pun. There are riding electric mowers but the cost keeps most off of them a…

The newest Zero electric motorcycles will go around 100 miles at real highway speeds, which isn't bad compared to most small-tank gas motorcycles (although obviously lack of refueling options is an issue). I do think they're a good illustration of the cost issue, though, as they're astronomically expensive.

Re: Tesla, software and disruption

#24

Post by justicezyx »

"When many car company people look at a Tesla, they see a not-great car with some cool features that they’re going to build too, being produced at a small fraction of the volumes they’re selling. “Look at the fit and finish, and the panel gaps, and the tent!” "

Is this the basis of the article?

The earliest Tesla car is rodster right? THAT IS A GREAT CAR in every aspect, right?...

Re: Tesla, software and disruption

#25

Post by mrfusion »

PAClearner wrote: post #13

for anyone who maybe got confused SLAM stands for Simultaneous Localization and Mapping. edit: I think if it is possible to do SLAM without LIDAR waymo will definitely beat TESLA to the punch since waymo has an absurd amount of vision data labelled with the lidar-measured-depth.

Speaking of that. I’d think we could possibly train on simulations. We’d have the 3d information plus the rendered 2d camera view.

Re: Tesla, software and disruption

#26

Post by frenchie4111 »

Something missing from this article: Tesla doesn't necessarily need to solve the vision only SLAM/Autonomy problem itself. If anyone solves the problem, and is willing to sell it. Tesla can immediately license it and turn their fleet of 200k+ Autopilot 2 cars into self driving cars overnight.

Re: Tesla, software and disruption

#28

Post by pcprincipal »

I read a lot of articles on TSLA and usually fail to see any mention of their power business. Granted, this is a small part of revenue right now ($1.1BN for FY17 of $11.8BN total), but growing fast and with some highly visible data points (Puerto Rico, South Australia). They are going to win multi-billion-dollar contracts if the US or other countries decide to rebuild their grids. As a TSLA shareholder, this opportunity actually gets me more excited about the future of the company than autonomy. It seems to me they have less competition doing huge battery installations than building autonomous vehicles.

Re: Tesla, software and disruption

#29

Post by PAClearner »

mrfusion wrote: post #25
PAClearner wrote: post #13

for anyone who maybe got confused SLAM stands for Simultaneous Localization and Mapping. edit: I think if it is possible to do SLAM without LIDAR waymo will definitely beat TESLA to the punch since waymo has an absurd amount of vision data labelled with the lidar-measured-depth.

Speaking of that. I’d think we could possibly train on simulations. We’d have the 3d information plus the rendered 2d camera view.

I don't know if people do this for vision yet but I know some people doing this for RL for robotic control.

Re: Tesla, software and disruption

#30

Post by Soundest »

frenchie4111 wrote: post #26

Something missing from this article: Tesla doesn't necessarily need to solve the vision only SLAM/Autonomy problem itself. If anyone solves the problem, and is willing to sell it. Tesla can immediately license it and turn their fleet of 200k+ Autopilot 2 cars into self driving cars overnight.

Whilst that's true, it doesn't seem likely that we go from where we are now (Can't do autonomous driving) to a point where can get fully autonomous driving, and then to a point where we can do fully autonomous driving with the reduced sensors available on existing Teslas in a period of time that makes sense.
Post reply on HN