Wow … so they are ready to go to 100k ?
It’s almost like they will give to anyone who shows some descipline.
How many FSD buyers are there in US anyway ?
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
FSD Beta 10.11
- Thread starter EVNow
- Start date
It’s almost like they will give to anyone who shows some descipline.
How many FSD buyers are there in US anyway ?
You have to remember Elon is very optimistic. Yes Elon may believe FSD will improve significantly with 10.11, and they will potentially double the invitees. More than likely, 10.11 will be another incremental improvement, and the plans of opening up to more users will change accordingly
Or it’s his way of giving people at least something for their $$. Plus think of how many more people will buy FSD, for now $12k, once they know the threshold for beta entry has changed to a very obtainable safety score.
tmoz
S85D, Y LR, ex Prius PiP
It'd be nice if those who paid for FSD, and are not yet allowed into the Beta program, could just have their credit cards on file and only have to actually pay when they are first allowed to check it out.
I have a hard time believing that would make a discernable difference (people using the lowering of the safety score for entry to beta as the basis for making a $12k investment). FSD is a long-term play and most of us bought in before there was a safety score or FSD beta.
I do agree with the notion of giving more people something for the money they've already paid.
EVJAutoJunkie
EV Junkie
I'm still baffled by the fact that Tesla does not have an official webpage on their website's member portal, that lists and tracks all these firmware releases.
Obviously you don't know Tesla very well!! To be expected.
sleepydoc
Well-Known Member
agree and disagree - this is consistent with Tesla, but I still would expect a company that is tech-centric and continually updating their software to have some sort of information on their web site so owners don't have to depend on 3rd party sites like notateslaapp.com
whoosierdaddy
FSD Supervisor
Why should they put work into this when the community will do it for free?
EVJAutoJunkie
EV Junkie
From a business standpoint, the last thing you want to happen is to have the public act as an auxiliary information outlet for the company's marketing department. Blogs and 3rd party websites are fraught with rumors and misleading information lead my misinformed users.
Surely, any sound marketing and legal team within a company would want to insure to the public that they are in control of their objectives and would not like to leave their narrative up to the peanut gallery.
For instance, I am pretty sure there are members who cite posts from this forum when disputing issues with Tesla's service department as if the service centers are obligated to comply. I wouldn't think that Tesla believes it makes their job easier.
I liked this so I decided to steal your data and go needlessly further with it.
first three columns are the same as yours. I just added days between 'major point' releases to filter out some of the 10.x.y hotfixes, and also added columns to compare that to the 'previous average', which essentially looks at the average days between release for all previous releases up to that point versus how long that specific release took.
edit: actually took a minute to look at the numbers now and with the exception of 10.11 they really have stuck to the 2-week release cycle pretty well. every 'point' release has been 2 weeks apart, +/- a couple days. 10.8 and 10.9 look weird, but really only because 10.8 was pushed early for christmas. If it had been 2 weeks after 10.7 like 'normal', then 10.9 would have been 2 weeks after that, as expected. with 10.11 releasing this week, it will have been about 6 weeks since the previous point release, which is the first 'break' of the 2 week cycle since at least september. This one has been a long wait, but sticking to 2-week release cycles for 5 straight months is fairly impressive
Last edited:
I think Elon has multiple incentives to open to more people
- The AP team is executing according to a plan and obviously the team leadership wants to show progress
- He feels compelled to open to more people because he "promised" he would
- Elon wants to put pressure on AP team
- To "show" progress, may be because he believes there is progress (based on data, though we may not feel it)
In any case, at this rate FSD Beta will go "wide" stealthily. When it goes from EA Beta to published "wide" there may not be much of a difference in # of people using it.
No, they have said that they won't recognize more revenue until it goes to wide release in a geographic area.
Yes - its likely they are on a 2 week sprint cycle. So, they will deliver every two weeks whatever they can. When we get those depends on whether there are any hot fixes that are needed.
The real break was 10.10.1. It came 2 weeks after 10.10 - probably because of that NHTSA recall. Now 10.11 comes 4 weeks after 10.10.1. So, between 10.10 that came on 2/1 to 3/15 there are 6 weeks and we are just now getting the next point release. So, hypothetically we didn't get the 2 internal releases in between.
Last edited:
I'm guessing they might start making the release schedule wider so that untrained beta testers have a bigger difference to report on.
So often the many "5% improvements" are very difficult to detect which gives the impression of no progress.
Some comments on the release notes.
Beta 11 features in 10.11 :
Beta 11 features in 10.11 :
- Upgraded modeling of lane geometry from dense rasters (“bag of points”) to an autoregressive decoder that directly predicts and connects “vector space” lanes point by point using a transformer neural network. This enables us to predict crossing lanes, allows computationally cheaper and less error prone post-processing, and paves the way for predicting many other signals and their relationships jointly and end-to-end.
- Use more accurate predictions of where vehicles are turning or merging to reduce unnecessary slowdowns for vehicles that will not cross our path.
- Improved the precision of VRU detections by 44.9%, dramatically reducing spurious false positive pedestrians and bicycles (especially around tar seams, skid marks, and rain drops). This was accomplished by increasing the data size of the next-gen autolabeler, training network parameters that were previously frozen, and modifying the network loss functions. We find that this decreases the incidence of VRU-related false slowdowns.
- Reduced vehicle “parked” attribute error rate by 17%, achieve increasing the dataset size by 14%. Also improved brake light accuracy.
- Improved clear-to-go scenario velocity error by 5% and highway scenario velocity error by 10%, achieved by tuning loss functions targeted at improving performance in difficult scenarios.
- Improved creeping profile with higher jerk when creeping starts and ends. (helps in other cases too)
- Improved smoothness through turns by using an optimization based approach to decide which road lines are irrelevant for c ? given lateral and longitudinal acceleration and jerk limits as w? vehicle kinematics.
- Reduced the predicted velocity error of very close-bymotorcycles, scooters, wheelchairs, and pedestrians by 63.6%. To do this, we introduced a new dataset of simulated adversarial high speed VI interactions. This update improves autopilot control around fast-moving and cutting-in VRUs.
- Improved control for nearby obstacles by predicting continuous distance to static geometry with the general static obstacle network
- Improved right-of-way understanding if the map is inaccurate or the car cannot follow the navigation. In particular, modeling intersection extents is now entirely based on network predictions and no longer uses map-based heuristics.
- Improved detection and control for open car doors.
- Improved stability of the FSD Ul visualizations by optimizing ethernet data transfer pipeline by 15%.
- Improved recall for vehicles directly behind ego, and improved precision for vehicle detection network.
That won't happen unless they go "wide" and remove the "early access" prefix.
And of course given the huge improvement in Tesla financials over the last year, it has become less important now. The annual report also expected less revenue to be recognized compared to earlier quarterly reports.
HookBill
Member
I have a suggestion for Elon. Instead of focusing on letting in those with 95 safety scores, why don’t you move those of us who paid for FSD back in late 2016, have made all the necessary HW upgrades and maintained a 100 safety score for quite some time, move to the top of the priority list. We are losing patience here.
Similar threads
- Article
