Until you meet their minimum view and subscriber levels, exactly zero dollars and zero cents.
MANY of the FSD beta testers fall below those limits currently. A fact conveniently overlooked by the thread FUDsters.
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
MANY of the FSD beta testers fall below those limits currently. A fact conveniently overlooked by the thread FUDsters.
Daniel in SD
(supervised)
Could be, however when you pay someone to do it you can fire them if they're doing a bad job. Another issue is that FSD beta testers seems to think it's fun to let the system drive erratically and/or disobey the law. I suspect paid testers are encouraged not to do that. Really I think it's too early to say, right now the system makes so many errors that only an insane person wouldn't pay attention.
BTW, it looks like a Cruise safety driver makes $25 an hour which admittedly is not that much in San Francisco.
Hyperchange is the worst because he's really peddling his own wares essentially.
He simply can't be non-biased.
Clearly there are lot of hand picked FSD testers that are not objective, and only handful of almost accidental aces like Chuck.
Chuck is a combination of a normal owner, balls of steel (who even signs up for taking 10+ uncontrolled lefts in a row), and someone with a lot of patience.
It's amusing to me how much discussion is around improvements vs. the most recent build. Of course there is going to noise and some videos are going to look worse and others better. The delta in improvement won't be big enough to make a clear difference in all testers.
What we should be doing is ask, how is this version vs. 3 months, 6 months, and 12 months ago?
It seems like there are clear significant improvements over the longer time frames, right?
Where do you think FSB beta will be in another 6 months, another year?
What we should be doing is ask, how is this version vs. 3 months, 6 months, and 12 months ago?
It seems like there are clear significant improvements over the longer time frames, right?
Where do you think FSB beta will be in another 6 months, another year?
Daniel in SD
(supervised)
I would guess that V10 is about 10x better than the first release. I predict (with huge uncertainty) in a year it will be 20x better than it is now.
It's a very difficult thing to quantify. I did a poll of V9.1 performance and the range of answers was 6 orders of magnitude apart!
Poll: FSD Beta V9.1, how close is it to release (i.e. out of beta)?
The stated goal for FSD is to be twice as safe as human drivers when used unsupervised. Tesla's safety data shows an accident rate of 1 per 2 million miles without Autopilot enabled (Tesla Vehicle Safety Report). They define an accident as "all crashes in which the crash alert indicated an...
teslamotorsclub.com
powertoold
Active Member
If Tesla *only* improved lane semantics and choice, V10 would be "10x" better than it is now. Add in improved pathing for complex and diagonal intersections, then you're looking at much much reduced disengagements.
What we're seeing with V10 is a great foundation to polish everything up. Obviously easier said than done, but Tesla is the only one with the data and data engine to get it done.
The logic here is: if Tesla was able to use their data to create such beautiful and accurate birds-eye-view visualizations, what's stopping them from extracting even more information from the cameras (perhaps other than hw3 limitations)? Tesla is closing in on LIDAR performance without any of the LIDAR drawbacks. During AI Day, Tesla already showed us the rich types of data that they can extract from cameras given the compute. They were basically proving that camera-only is viable.
Last edited:
Crazy how she's getting attacked. Shows you how tesla fans are the worst breed of fan. Literally have never seen anything like it.
Daniel in SD
(supervised)
Chuck gets wayyy too much credit. Him repeatedly testing the exact same insane left turn seems unnecessary and dangerous. Once, maybe twice is sufficient. We get it, the current iteration cannot reliably complete the insanely difficult left turn. I can’t imagine they need him to test it 25 times in a row.
Move along and drive as if you’re going about a more typical routine with a huge variety of different lefts, rights, rights in red, yellows, etc. Far more likely to catch various edge cases and video useful for the NN and mothership.
Ofcourse its gonna improve. The real question is what is the rate of improvement versus other AV systems. But based on what we are seeing their rate of improvement is akin Nissan and not Google. They have a very average but subpar rate of improvement.
Tesla fans always claim this exponential magic rate that is 100% fabricated and doesn't exist. What tesla fans should be praying that their rate of improvement is atleast as good as Waymo.
Waymo
2009 - 25 mile (based on their 10 100 mile uninterrupted challenge)
2012 - 500 (guesstimate)
2015- 1,244.37 mile (CA Disengagement report)
2016 - 5,127.97 mile (CA Disengagement report)
2017 - 5,595.95 mile (CA Disengagement report)
2018 -11,154.3 mile (CA Disengagement report)
2019 - 13,219.4 mile (CA Disengagement report)
2020 - 29,944.69 mile (CA Disengagement report)
2021 - 35,000 (projected based on history)
2022 - 90,000 (projected based on history)
Last edited:
@Daniel in SD wait does vegas have a line on that? A bit ...edgy to have a pool on this but...somehow enticing.
I kid I kid
I kid I kid
powertoold
Active Member
Interesting situation, grey area for me, in terms of what V10 did and how confidently it did it. But yes, I agree that Whole Mars should have been more ready to take over, although having hands on the wheel or not doesn't equate to being more ready IMO. I find that I'm more ready to take over AP when my hands are close to the wheel but not on them. This seems counterintuitive until you've used AP for a while. Having my hands on the wheel turns my reactive brain off. Plus, there's not much he could have done there in terms of turning or stopping for oncoming traffic.
stopcrazypp
Well-Known Member
Yes, I pointed this out a long time ago in the first video that the car tried to go between the pillars. That apparently is a legal move.In WA where the video was shot, changing lanes across the solid white is apparently legal
Can you slalom through Seattle's monorail columns on 5th Avenue?
A recent on-air debate about driving back-and-forth between the concrete columns under Seattle's monorail spurred a short investigation, and a shocking traffic revelation.mynorthwest.com
Note however the cop they initially asked thought it wasn't legal but said they need to check with Seattle DOT to be sure. DOT confirmed it WAS legal.
stopcrazypp
Well-Known Member
Yep, nothing it the CVC prohibits crossing a single white line. There is mention in a handbook that is it discouraged, but nothing says specifically that it is illegal. In most situations it is illegal due to other signs that take precedent, but not the specific act of crossing the line.
Can you cross a solid white line on a California road?
There is a carpool lane in a bridge that I cross all the time. The entire lane is separated by a single white solid line and there are no dashed breaks anywhere along the whole way (not even at the start where it begins at a concrete barrier). If it was illegal to cross a single white solid line then that carpool lane would have effectively been off limits to everyone, given you must cross it at some point to enter.
EVNow
Well-Known Member
Very, very little - if anything at all.
Most you tubers survive on sponsorship.
EVNow
Well-Known Member
The guy who I asked it to isn’t answering because he knows he is wrong.Very, very little - if anything at all.
Most you tubers survive on sponsorship.
stopcrazypp
Well-Known Member
I should point out even if people made lots of money from Youtube videos, a key point is it's not predicated on saying positive things (about Tesla or otherwise). So people can say completely negative things about Tesla (there are some channels dedicated to this) and still make money if they go past the monetization line. I think there's enough FSD Beta videos out there that we know Tesla isn't really filtering things, as we get a look at all aspects (even when it is performing quite poorly). It's a view we don't really get from other ADAS development.
I fully admit my info was outdated, I didn't realize how severely youtube has cut payouts since my content creation days.
I still don't like Tesla playing favorites but I'm less offended by it now.
No, the guy you asked was doing his job
Unlike Elon, I'm fully capable of admitting when I'm wrong.Similar threads
