Thread by @OgMage78

I got #FSDBeta on Christmas Eve 2021. (Thanks @elonmusk & @Tesla) I got it originally on our 2018 Model 3 and then again when we traded Tess in for our new 2022 Model Y (Y Not). I am not on the new 10.69 version yet, (1/26)

so all these observations will be on overall progress from late 2021 to now and the current version I am on 2022.12.3.20 I am a super nerd, (2/26)

so I have purchased FSD twice to be able to watch the progress in real time and to hopefully provide useful feedback via the appropriate use of send snapshots to Tesla and my interventions.
The progress between releases is obvious. (3/26)

As they iterate over different solutions, you get the typical 2 steps forward, one step back behavior you would expect in any beta program. The number of interventions, (4/26)

the smoothness of the drive and the predictability of how the car will behave have all improved over the last 8 months.
My personal use: My wife and I both use the car so when I drive, I turn on FSD about 90% of the time. To and from work to the store, (5/26)

around town just running errands, any trips out of town – if I am behind the wheel of the car, FSD is driving.
Observations:
100% of my drives to work (about 4 miles, well-marked roads with no construction, moderate complexity) no interventions. (6/26)

That was not the case several months ago, there is a weird entrance to the office, and it would get confused. That is no longer the case.
Any drive local or on road trips either in downtown Houston, or in Podunk small towns, (7/26)

if the road markings are clear and there is no crazy construction, I would say FSD is rock solid 99.9% of the time. It is a very rare intervention. This gives me to most optimism about the software and its potential. I have taken my car to Austin, Texas, Natchez, (8/26)

Mississippi (very old town with tons of one-way narrow streets and antiquated street design), Oklahoma City, Oklahoma, and lots of small towns on the trip from Houston up to north Louisiana. Again, (9/26)

barring torrential rain when I can’t see to drive or crazy road construction, FSD performs exceptionally well. Very rare for an intervention.
Road Construction – If it is light construction, cones, merge lane closure with cones – not bad. (10/26)

It handles those situations 95% of the time. Old confliction lane markings, lots of switching lanes with misplaced cones – not great yet.
Heavy construction, I let it try, but generally intervene and I stay hyper aware of the car in those situations. (11/26)

There is a super charger outside of Waco, Texas in the parking lot of Collin Street Bakery. The construction there is brutal. Roads, blocked, massive diversion for construction, temporary ramps – FSD will route me in a circle trying to get there. I just snapshot it, (12/26)

turn it off and manually drive to the super charger.
Bad weather – It can see (via the visualization) much better than I can as visibility degrades. I don’t know what the plan will be in the end state of FSD. (13/26)

There are times when humans must pull over and wait for the torrential rain to clear. I am not sure, at Level 5, what the path forward is. I don’t see how a vision-based system can work if your visibility is zero or near zero. (14/26)

Is the plan to have the car safely pull over as a human would do? @aelluswamy or @elonmusk would have to answer that question.
Overall assessment – the software is very impressive. There is an old saying, it is not how well the bear dances, (15/26)

but that it can dance at all. This product will have life changing effects once it is feature complete. The time saved, the opportunities for visually impaired or older drivers to maintain some independence – it is mind boggling to think about. (16/26)

I think Tesla has an outstanding team on this, and their approach is going to end up being the correct choice. To paraphrase what Elon has said many times, our transportation infrastructure was designed to be used by two low quality cameras on a swivel. (17/26)

What gives me concern? Just the randomness of our world. It is going to be a long march of 9’s to get to where you can tell your car where to go and then take a nap or watch a movie. (18/26)

The rollout worldwide to different countries with their varied driving rules is also another issue I wonder about the timing of ultimately.
I was quoted in an Arstechnica article back in October of 2018. “Elon Musk always delivers, (19/26)

but his deadlines aren't always the most accurate," said Charles Scott, who lives in the Houston Area and bought a Model 3 this summer. When it comes to delivering on his promises, Scott says, Musk operates on "Elon time. (20/26)

"
Scott says he knows full well that he's "not going to be taking a nap in the back of the car in the next 6 months." He expects full self-driving capabilities to come out in two to three years. (21/26)

And he expects that those who pre-paid for the full self-driving capabilities will get any necessary hardware upgrades for free.”

I still see that as the most likely outcome. (22/26)

I probably thought in 2018 it was a 3-to-5-year time frame to FSD being feature complete (Not counting governmental approval). That would put us in 2023. Based on my current testing, that might be a little optimistic. (23/26)

the question is the timing. I am as excited to watch the continued evolution of the software as I was back in 2018. It’s capabilities now vs 2018 is to borrow a phrase from @WholeMarsBlog “mind blowing”. I hope this has been useful. (25/26)

Here is a link to the Arstechnica article for proper attribution to the writer.
arstechnica.com/cars/2018/10/t…

Chares

(26/26)

@threadreaderapp Please unroll this lengthy tale