Search results

Alright @Trent Eady put your money where your mouth is. Good luck, champ. Trent Eady’s Misuse of Baidu’s Research

I wish it were so simple. When I have challenged his conjectures here, on seekingalpha, etc. he just begs a moderator to protect him, to enable him, by having the challenging posts removed. He does not want open discussion — he wants an adoring audience, and an environment he can control. This...

As I said, the problem is that he freely mixes facts with his own conjectures. It’s very difficult to tell them apart, and he makes no effort to distinguish them. Further, I think he actually intends to present his conjectures as facts. He craves an audience more than anything in the world, and...

He mostly just copies and pastes stuff he finds on other websites. When he editorializes, he always gets a bunch of stuff wrong. Sometimes it’s subtle, but it’s often obvious to a knowledgeable reader. This is a problem. He spreads lots of misinformation mixed in with his news recaps. It can...

There is more to this. What do you want this forum to be? Are you ok with it falling into de facto control of an obvious crackpot, who is actually using you for your moderator privileges? Look around — many people here would rather @Trent Eady go away.

Hey mods — you get to make a choice today about one of the most prolific crackpots ever to post here. @Trent Eady fulfills every part of the crackpot rubric: - He has an overarching “thesis” that he pushes with every post. The thesis is tenuous, reductive, simplistic, and almost universally...

@Trent Eady why don’t you make your own competing forum? You could control the whole thing like a lord over peasants. Everyone loves hearing you talk, and your understanding of ImageNet is so deep and valuable that I’m sure people would flock to your forum!

I know, right??? They should be more like Tesla and just say something really honest, like: “In the 3rd quarter, we registered one accident for every 4.34 million miles driven in which drivers had Autopilot engaged... By comparison, NHTSA’s most recent data shows that in the United States there...

Please, please, please make @Trent Eady moderator. You just have to look at the wild success of the competing forum he runs. Check out my screenshot! Now imagine how cool it would be if every post on TMC is just @Trent Eady speaking his mind the same way. He is clearly an expert and leader in...

Yes! This forum contains ideas and opinions that are contrary to the clear truth that “more data solves everything lol.” I mean, just look at how ImageNet performance scales with more training data! It’s clear as day that self-driving is just like ImageNet, so we have to stamp these contrary...

Also it’s so thrilling to see that @Trent Eady has really taken the Baidu paper I showed him ten months ago to heart! :) He actually includes it in all his articles now, because it is so scholarly. It’s okay that he cherry-picks one figure from it and ignores the rest! That’s how you do science...

Yes! This forum contains ideas and opinions that are contrary to the clear truth that “more data solves everything lol.” I mean, just look at how ImageNet performance scales with more training data! It’s clear as day that self-driving is just like ImageNet, so we have to stamp these contrary...

Yes! Tesla is the only software company that has a development version that is more capable than their production version. They pioneered this form of software development, and it is just a matter of time before it all pays off. Those extra 9’s of reliability are each just a single order of...

Elon hired Karpathy after showing him *twelve* different breeds of dogs — and Karpathy guessed all *ten* right. Turns out the joke was on Elon, and there are really only ten breeds of dogs! Who knew, except Karpathy?? That’s why Karpathy was INARGUABLY the best guy to hire to create a...

Tesla’s strategy is absolutely to wait until their FSD development branch is totally reliable, essentially incapable of any kind of failure, before doing an OTA to wake up the fleet of robotaxis. Knife-switch releases are clearly the most prudent and safe way to release software updates. Just...

Absolutely! Radar is expensive. That’s why Tesla only uses a single, low-quality radar. It’s clearly completely adequate for seeing things like police cruisers. Tesla on Autopilot crashes into police car

Right! Google doesn’t know anything about hardware. They are just a bunch of clueless software engineers who just happen to have the largest and most sophisticated datacenters on the planet. I’m pretty sure they bought it all from Dell. Their hardware has so little impact that it only handles...

You are right! Tesla is the only company that has ever used techniques that have been widely publicized in CVPR and NeurIPS papers with thousands of citations each.

Yeah, I mean, it took Waymo ten years, so Elon could probably do it in like thirty, fourty-five minutes. He’s just waiting for a dominant competitor to prove it’s possible first, before even trying. That’s just good business sense.

It’s good to see that you still think literally anything is always good news for Tesla. The simple truth is that Tesla is going to win every race, no matter what competitors do. Just listen to Automomy Day if you don’t believe me. They even showed figures from Google papers to prove their...

No, this is not possible. Training a net means showing it millions of training examples, often images. Images are large in size (bytes). A training set can be many terabytes in size, and they might have many such training sets. Common household network connections are orders of magnitude too...

This conversation is boring now. C’mon, won’t someone say something ridiculous? @CarlK where are you? :(

It just makes it all the more sad that, six months later, they’re depending upon their customers to draw lane lines in their local parking lots. It’s almost like they were just giving a marketing pitch.

NOPE. If they set their threshold all the way down to 1% confidence, then they will miss all the unfortunate objects that only had 0.5% confidence. 99% recall is nowhere near good enough for a safe AV. And the 1% they’d miss? Well, those would be the hardest examples... the ones they’d need the...

Agreed, those are also part of an HD map. But do you see that Tesla now uses maps that have hand-edited lane lanes? Their maps now also include traffic lights and overpasses. It’s a very slippery slope, my friend. Tesla is the map-less company that just keeps adding more and more dependence on...

Yes! Lanes are features of an HD map. The people in this thread are talking about editing OpenStreetMap to add lanes and obstacles and such to make Smart Summon work. To make it work at all, in fact. After all this bullshit about map-less driving, it turns out that Smart Summon uses an HD map...

What a fascinating non-answer. I guess you’re a not a champ after all. I gave you too much credit. Sad trombone. I’ll ask again: what specific features do you think makes a map “HD?” What exactly makes it... “super well maped...” vs. just “regular old maped?”

Ok, champ, tell our viewers what exactly you believe qualifies a map as “HD?”

The arguments that Tesla fans make against maps are: 1) The effort required to make a map is staggering and insurmountable. 2) Construction, police activity, etc. constantly change the state of the world, and no mapmaker could keep up. Neither of these arguments make a distinction between 2D...

Wait, I thought Tesla was the pioneer of map-less driving? But it actually depends heavily on OpenStreetMaps????? Hahahaha. That is ******* hysterical to me! This is a very good observation, and one of the most potent arguments against "fleet learning." False negatives (FNs) are among the...

Everyone — literally everyone — uses deep learning for vision. Zoox, Cruise, Yandex, everyone. Literally everyone. Waymo can even read police officer hand gestures.

Literally every AV company uses deep learning for vision. I think you may be misunderstanding the situation here. There are several people here who know a great deal more about this subject than what you could learn from watching the Autonomy Day presentation. That presentation was...

So wait... you don’t understand what I’m saying, so *I* need to go learn something? Deep nets are retrained all the time with new and better data sets. Each iteration is brand new. There is no carry-over of learning from one training run to the next. Every time you retrain a net, you risk...

Where on earth did you hear these things? From PopSci: “In fact, parking lots are a distinctive enough environment that Waymo, the self-driving car company that’s a sibling to Google, specifically trains its vehicles to deal with them by setting up real-world scenarios in a controlled...

I am fascinated by his lack of real knowledge or understanding, coupled with his absolute confidence in his beliefs. As my wife would say, "all of the wrong people have high self-esteem."

This is a misconception. A deep net has a capacity that is a function of the number of weights. The number of weights is constrained by the amount of compute available on the car. Today’s deep nets have ~100M weights. Larger nets take too long to run, and cars need quick reaction times. A net...

Then why don’t they do anything autonomously, even now, almost four years later? Literally every AV company uses deep learning. For ****’s sake, the Udacity Self-Driving Car nanodegree focuses on deep learning! Do you understand how common something has to be for it to get baked into an...

Fair enough. I only got into this to call attention to the insane bullshit posted by @CarlK. But then y’all kept piling up on me anyway.

Right, Waymo has the ability to remotely direct its vehicles. It does not simply remotely control them 1:1 with human drivers in sweatshops, or whatever the hell else has been argued here. That would be stupid.

Latency, unreliable connectivity, etc.

Why hasn’t anyone tried to address my questions?

I’m an authority on the subject.

Dude, the person I was responding to, before you got here, was @CarlK. The things he said were indeed very stupid. He really does believe that Waymo is just remote controlling cars, 1:1, despite how laughably ridiculous, pointless, and infeasible that would be. And if the car is remotely...

So Tesla fans now believe that Waymo just made a remote controlled car and is just lying about achieving full autonomy. I have a couple questions. 1) If it’s just remote control, why did they choose to do it Phoenix instead of in a harder place with more potential, like NYC? I mean, if it’s...

Yes, it is. AV hardware is very expensive. It makes no sense, economically, to drive cars remotely. You’d be better off just manning the car. Remote operation can’t be made safe, either. No one is doing that, because it would be stupid. Uh.... do you think the car needs to be monitored while...

CarlK literally said “remotely monitoring each and every car so they could apply necessary actions when there is a need.” That’s 1:1, which would be stupid. *clutches pearls*

Why would a 1:1 remote driver : car ratio be stupid? Seriously?

No, that’s not how Waymo cars work. That would be stupid.

No, that’s not how Waymo cars work. That would be stupid.