Lyft releases self-driving research dataset • Hacker News Board

Re: Lyft releases self-driving research dataset

#51

Post by e40 » Wed, Jul 24, 2019, 4:59 AM UTC

malandrew wrote: post #36

rubyfan wrote: post #30

Earlier quoted context omitted.

If you are an officer of scale you should take this offline. If not then you’re probably not authorized to speak on behalf of scale. Check your NDAs and service agreements. This is in such poor taste, anyone who would consider scale’s service now has to consider this sort of public commentary.

Leaving out attribution is in poor taste. There were multiple chances to give attribution in the original blog post.

We don't know the terms of the deal between the two companies. Your comment makes a lot of assumptions.

Re: Lyft releases self-driving research dataset

#52

Post by FBISurveillance » Wed, Jul 24, 2019, 5:00 AM UTC

ayw wrote: post #38

cardigan wrote: post #13

(I work at scale) Hmm this blog post and the website doesn't mention that this dataset was mostly annotated by Scale (scale.ai), as part of a partnership with Lyft ... We're going to publish a blog post about this soon, but if anyone at Lyft is reading this, please figure out how to reasonably credit Scale since I doubt leaving out Scale completely from the announcement is in the spirit of the agreement. Scale should…

Hi, I'm the CEO of Scale.ai. This comment does not represent the company's viewpoint, and cardigan is not speaking on behalf of Scale. We are very excited to have been able to work with Lyft in open-sourcing this dataset and advancing the research community. We are also very grateful to Lyft for choosing to leverage our point cloud viewer and have credited the annotations to us on their launch page.

That escalated quickly.

Re: Lyft releases self-driving research dataset

#53

Post by exit » Wed, Jul 24, 2019, 5:12 AM UTC

obese_by_nature wrote: post #32

cm2012 wrote: post #22

Earlier quoted context omitted.

Quick tip: voice concerns about partners in private. Lyft will probably be happy to credit Scale more - it was likely an honest mistake. But now you dragged them through the mud publicly, which is going to make big companies less likely to work with Scale in the future.

[flagged]

[flagged]

Re: Lyft releases self-driving research dataset

#54

Post by kybishop » Wed, Jul 24, 2019, 5:21 AM UTC

posix_compliant wrote: post #21

cardigan wrote: post #15

Earlier quoted context omitted.

Also, the viewer packaged with nuScenes was built by Steven Hao from Scale, and while it was packaged as part of nuScenes it should probably be called Scale's viewer instead of nuScenes' viewer. The original viewer in the nuscenes SDK has the Scale logo, but it looks like Lyft removed that in the fork. Maybe a bit of public shaming will fix that... Dear Lyft marketing person who wrote this: we are a data labeling com…

Not nitpicking, genuinely curious. L10? I've never heard of that climbing scale.

There is no such thing. They have to have meant v10, and climbing v10 "in a few months" is an incredibly skeptical claim.

Without some kind of gymnast background I honestly don't think it's possible. Even 0 to v10 in a year is hard to believe. I've heard of phenoms doing it in 2 or 3 years, and don't doubt a year is possible... but a few months? Kind of like going from couch to sub 5-minute mile in a few months.

Re: Lyft releases self-driving research dataset

#55

Post by dangero » Wed, Jul 24, 2019, 5:31 AM UTC

exit wrote: post #50

ayw wrote: post #38

Earlier quoted context omitted.

Hi, I'm the CEO of Scale.ai. This comment does not represent the company's viewpoint, and cardigan is not speaking on behalf of Scale. We are very excited to have been able to work with Lyft in open-sourcing this dataset and advancing the research community. We are also very grateful to Lyft for choosing to leverage our point cloud viewer and have credited the annotations to us on their launch page.

the point cloud viewer which cardigan also had to point out was produced by someone at scale, and not nuScene? pretty inconvenient for you when workers want their efforts recognised, and not just thought of as capital for you to push around however you please, huh?

That’s a strange take on this. Not all staff are authorized to speak on the company’s behalf. That’s true almost anywhere I’ve worked. Your efforts cannot always be recognized externally. NDAs and various other types of contracts commonly outline that.

Re: Lyft releases self-driving research dataset

#56

Post by 1000units » Wed, Jul 24, 2019, 5:34 AM UTC

mitfahrener wrote: post #41

cardigan wrote: post #15

Earlier quoted context omitted.

Also, the viewer packaged with nuScenes was built by Steven Hao from Scale, and while it was packaged as part of nuScenes it should probably be called Scale's viewer instead of nuScenes' viewer. The original viewer in the nuscenes SDK has the Scale logo, but it looks like Lyft removed that in the fork. Maybe a bit of public shaming will fix that... Dear Lyft marketing person who wrote this: we are a data labeling com…

Have a wonderful day tomorrow. Good luck.

I'm hoping this was coordinated with the CEO and he's getting free lunch tomorrow for the successful publicity stunt.

Re: Lyft releases self-driving research dataset

#57

Post by 1000units » Wed, Jul 24, 2019, 5:36 AM UTC

goguy wrote: post #42

cardigan wrote: post #15

Earlier quoted context omitted.

Also, the viewer packaged with nuScenes was built by Steven Hao from Scale, and while it was packaged as part of nuScenes it should probably be called Scale's viewer instead of nuScenes' viewer. The original viewer in the nuscenes SDK has the Scale logo, but it looks like Lyft removed that in the fork. Maybe a bit of public shaming will fix that... Dear Lyft marketing person who wrote this: we are a data labeling com…

"Gets to number one in random video games" Genius

Intelligence makes you good at everything, it's almost not fair.

Re: Lyft releases self-driving research dataset

#58

Post by mkagenius » Wed, Jul 24, 2019, 5:39 AM UTC

ayw wrote: post #38

cardigan wrote: post #13

(I work at scale) Hmm this blog post and the website doesn't mention that this dataset was mostly annotated by Scale (scale.ai), as part of a partnership with Lyft ... We're going to publish a blog post about this soon, but if anyone at Lyft is reading this, please figure out how to reasonably credit Scale since I doubt leaving out Scale completely from the announcement is in the spirit of the agreement. Scale should…

Hi, I'm the CEO of Scale.ai. This comment does not represent the company's viewpoint, and cardigan is not speaking on behalf of Scale. We are very excited to have been able to work with Lyft in open-sourcing this dataset and advancing the research community. We are also very grateful to Lyft for choosing to leverage our point cloud viewer and have credited the annotations to us on their launch page.

Business should not be lost. Credit is secondary.

Re: Lyft releases self-driving research dataset

#59

Post by mkagenius » Wed, Jul 24, 2019, 5:44 AM UTC

bearpelican wrote: post #49

jacobn wrote: post #8

“There will be $25,000 in prizes, and we’ll be flying the top researchers to the NeurIPS Conference in December, as well as allowing the winners to interview with our team.” I guess it’s a decent opportunity if you’re trying to break into DL?

I had a bad experience with Lyft's previous self driving challenge - https://www.udacity.com/lyft-challenge I unofficially got first place after finding a bug in their test set (allowing me to blow the competition away). I reported the problem directly - they decided not to fix it and asked me to take my submission down. They said they'd still offer an interview. However - this interview wasn't even for their DL team…

> they decided not to fix it and asked me to take my submission down.

Imagine if it was a bug in their app leaking millions of users data and they go "We do not fix it. You can only use Uber now" "Also, come for an interview if you want to fix our web tools instead"

Re: Lyft releases self-driving research dataset

#60

Post by mkagenius » Wed, Jul 24, 2019, 5:49 AM UTC

arathore wrote: post #9

azinman2 wrote: post #6

$25k in prizes seems silly given this is a multi-billion dollar market to crack.

Such competitions do not usually result in a comprehensive "solution" by themselves - pushing the state-of-the-art is more common. Also the value is not going to be derived solely from the algorithm but more from its deployment to real world applications and the surrounding infrastructure to make it possible.

> pushing the state-of-the-art is more common

But do not forget there will be 10s (if not 100s) of people working on this for 30 days. The man-hour this competition will use is highly disproportionate to the amount offered overall.