More

    Review: Tesla Full Self Driving (June 2020). Should you buy FSD before July 1st ?

    There’s just a few weeks left to make your decision on buying Tesla’s FSD package or not, ahead of a price increase on July 1st.

    After having saved hard and forked out the cash myself on May 2nd, I’ve now had time to live with the FSD features and are ready to break down my thoughts on the upgrade.

    The Price

    Let’s start with the price of the upgrade. Currently, the FSD upgrade will cost you A$8,500 and Elon has indicated that this will increase by around $1,000 at the end of the month. In Australia that is likely to be at least A$1,100 if they’re not including GST in that price. If it’s a literal price translation of USD to AUD, then we could be talking about a price rise as much as A$1,500.

    At the current price, it’s definitely a significant investment and after the price rise, could be as much as A$10,000. Given you could buy some pretty decent second-hand cars for that price, the features clearly need to offer substantial functionality to be justified.

    Your ability to find the money to purchase FSD is also dependent on your financial position, regardless of how good the FSD package is. Many Model 3 owners spent more on buying the car than they ever have before. This financial stretch to achieve the purchase, means you could find it difficult to find another 10k on top.

    If you find yourself in this camp, it may be tempting to wait till the end of the year when Tesla plans to launch FSD as a subscription. Just understand we don’t know the price per month, but it won’t be cheap. By the time the price is revealed, the option to get unlock FSD for A$8,500 will be long gone.

    Musk has often suggested that the value of your Tesla will increase over time and recently suggested the value increase could be as much as $100,000.

    For those numbers to be a reality, Musk is clearly including potential income created by adding your vehicle to the Tesla Fleet. This would enable you to pay back the purchase price of FSD quickly, but not everyone will be prepared to do this with their pride and joy.

    If you do decide to purchase, the software unlock is almost instantaneous now. While the early days it could take a day or so to activate, clicking buy on the upgrade, will see the new capabilities arrive in your car in just seconds.

    Now for a look at what you’re buying with the FSD package.

    Smart Summon

    Originally summon started out as a simple feature that allowed you to move your Tesla forward or backwards. This helped get your car out of a tight garage, or if you found other drivers had parked too close to enter your car.

    Now upgraded to Smart Summon, Tesla now promotes this feature as ‘your parked car will come find you anywhere in a car park.’ While that’s certainly the ambition and I actually have no doubt it will in the future, right now that’s not exactly true.

    The feature is in development, which means the car often suffers from local routing issues, frequent pausing and generally drives slower than humans would when navigating a carpark.

    The goal of Tesla’s FSD technology is to surpass human ability, but so far, Smart Summon fails in this regard.

    Now for the positive. It is bloody amazing that a $100,000 car can basically become a giant RC car that can move, without anyone in it. There are exactly zero other cars on the market that offer this. For that reason alone, it really is impressive.

    I know many other Tesla owners try this feature once or twice to show off to friends and then never use it again. Personally I’ve made a conscious effort to continue using it every few days, or basically any time I have the opportunity to (in a carpark).

    I’m happy to report there is being progress being made with Smart Summon. While not detailed in the release notes of software updates, the system is clearly learning and improving.

    My biggest complaint initially was the routing, as it completely ignored the white parking bays on the ground, something humans would typically avoid. Now it seems to be taking more efficient routes and getting better. It’s definitely not perfect, like the limitation of only being able to deal with 10 degree inclines or declines making it unusable for split-level or multi-story car parks.

    I really look forward to the day where its pouring rain and I can summon the car to come to pick me up. A drop-off and pickup near the entrance of a shopping center or your workplace, would almost entirely eliminate the need to carry an umbrella. It would also be a very clear differentiator between Tesla and all other vehicles. I’m not sure what I’d pay for that specifically, but let’s call it a lot.

    After having successfully used Summon a few times in the past week, I have noticed there’s a very first world problem. If you walk to your car, you may pop the frunk and hook your groceries over the hooks, close the hood, get in and drive away.

    If your car comes to pick you up, you’re much more likely to get in the driver’s seat carrying whatever you have in your hands, especially if there’s traffic around. As I said, it’s a first-world problem, but not one I think many have considered.

    HomeLink

    If you got one of the early Model 3’s sold in Australia, then you’ve likely had your HomeLink hardware retrofitted after delivery. If you use the car’s Summon feature to reverse your car out of the garage, it’s smart enough to fire a command via HomeLink, to open your garage door, before reversing.

    Personally my driveway is on a decline, so it gets 3/4 of the way out, then stops. Again, I hope this limitation is solvable with the current hardware suite and is simply a software problem.

    Another aspect of HomeLink that unlocks when you purchase FSD, is the ability to open and close your garage door from a new button in the Tesla Mobile app. There’s not really any reason this should be restricted to FSD customers, but that’s how Tesla are currently structuring the feature.

    I actually really like this feature, but would love to see the control moved to the home screen, rather than located under the Controls menu.

    Navigate on Autopilot

    This is probably the headline feature of the FSD package, offering the ability for the car to drive on-ramp to off-ramp, indicating for you and automatically overtaking slower cars.

    NoA is definitely a great feature and actually solves a real issue on long drives. Before FSD, I was frustrated when I’d encounter a slow vehicle on the freeway, having to indicate, make the lane change, pass then, then indicate and return to the left lane. Then repeat that a couple dozen times.

    The system still isn’t perfect, but again it is improving with more Australian data. I noticed that on a freeway entrance the system would turn on the left indicator, for a right merge onto the freeway. I also found it’d lane change me to the right lane, with no other cars around.

    This felt like some of the code from the US was still present and not flipped to reflect our RHD market. Thankfully after a couple of OTA software updates, most of these issues have been resolved, but your experience may vary based on your location.

    One of the choices you have to make when enabling this feature is how aggressive you want the lane changes to be. In an effort to test the best/worst-case scenario, I immediately went for Mad Max mode. Even at this setting, I found the car left an acceptable distance to the car it was overtaking, but it is good to have less aggressive features to provide comfort to drivers.

    It is clear with NoA, that we’re on a path where these cars can drive themselves. This is perhaps the best demonstration of the car’s understanding of its surroundings and ability to safely navigate through the environment. This enables you to drive (observe while the car drives) over much longer distances and get out at your destination having far less mental fatigue than any other vehicle.

    I do get frustrated by having to provide force to the steering wheel and would love it if Tesla would enable driver monitoring instead. Personally I use Autopilot by holding a little pressure in the downward direction with my right hand, it’s comfortable and sustainable on a long journey with my elbow on the armrest.

    When the car needs to move left to complete an overtake, it’s a weird sensation to have the car pulling in the opposite direction, almost fighting me. I have to provide just enough pressure to let it know I’m paying attention, but not enough to cancel the lane change.

    Lane changes on Autopilot

    This is really an undocumented feature of the FSD package and one of the best surprises with the purchase. When using regular Autopilot on a dual-lane stretch of road, you simply flick on the indicator, the car makes the assessment on surrounding traffic and when clear, will make the lane change for you. Even better still, it also disengages the blinker too, this really is nicely done.

    To be clear, this is not using Navigate on Autopilot, which means you don’t need a destination defined. I have an 80km/hr stretch of road on my commute and this feature makes driving much easier, making it easily one of my favourite FSD features, used daily.

    This feature works so well that it basically removes the need to do head checks and also makes blind spots a thing of the past as the external cameras and sensors don’t have blind spots like the A, B and C pillars that drivers have.

    Autopark

    After having FSD for just over a month, I’ve definitely used the Autopark feature the least, maybe twice. This is simply because the Autopark option appears so rarely. While Autopark can park in both parallel and perpendicular spaces, I often try to park away from other cars, avoiding door dints.

    I know some people are terrible at parking and this really could be a substantial feature for them. The disappointing thing is that its ability to assist you is based on having other cars around you. Tesla should be able to use the line-markings on the ground to guide the car, even if there are no adjacent vehicles.

    Once other aspects of FSD like Smart Summon get better, I’ll have to come to terms with the fact I’m not choosing the parking spots anymore. For this reason I think Autopark is a feature that’ll be leveraged more by the car in the future, instead of me right now.

    Upcoming

    Still on the to-do-list on the website are the following two items:

    • Recognise and respond to traffic lights and stop signs.
    • Automatic driving on city streets.

    The first one is actually half way done. In software update 202.12.11.1 we seen the on-screen visualisations updated to show traffic lights, road markings and even wheelie bins. The second part of the feature involves responding to a red light or stop signs, something already available (in beta) in the US.

    Musk has confirmed the Traffic Light and Stop Sign Control (beta) is being expanded to Canada this Thursday, so we’re hoping Australia and the rest of the world isn’t far off.

    This update enables the car to recognise it is approaching an intersection and asks the driver to confirm they want to proceed through (if the light is green). This is training the system to accurately identify what humans do given the current situation and with enough data, the confirmation can be removed.

    This is an important aspect of the car being able to drive itself without human intervention, as is the last feature on the FSD to-do-list.

    Finally, we come to one of the hardest outstanding challenges – City streets. There is a serious amount of learning required for a car to successfully navigate the millions of complexities that occur at medium speeds.

    There are challenges like slowing for speed bumps, roundabouts, island dividers, hook turns and dealing with a lack of lane markings. Also there’s a far higher chance of having to navigate pedestrians, bike riders, double-parked cars and emergency vehicles.

    Final thoughts

    What Tesla is doing here is training a computer (FSD) to be like a human brain, powered by data, it needs super-human memory and recall ability along with eyes that see in every direction.

    Musk has said FSD will be feature complete by the end of the year (he also said that about 2019), but if that can be achieved, the car is likely to be able to transit from most people’s garage, to their work, without intervention.

    FSD won’t be perfect for quite some time, which means we’re unlikely to get to text and drive, have beers or go to sleep for at least a few years.

    Having your daily commute driven for you, will be a massive step forward, even if the solution isn’t perfect in all scenarios for a while and something that’ll be enough of a justification for many to buy the FSD upgrade.

    A purchase of FSD today is absolutely an investment in the medium to long-term future, while buying now locks in the price and ensures you aren’t impacted by future price rises.

    I bought FSD now because I believe Tesla will be the first in the world to have cars that can drive themselves. FSD is already an impressive lineup of features, and I’ll enjoy getting to experience the future develop.

    Personally, I understand how Tesla is using AI to teach that brain, powered by their growing fleet of vehicles, collecting data from the millions of km completed across the world. Having designed amazing hardware, this is now a software and data problem and the progress to date suggests the target will be achieved, albeit a little later than we’d hoped.

    Other Tesla owners will be keener to wait until the system is done and then buy-in. That’s also a perfectly acceptable approach, but you have to understand FSD is likely to be much more expensive by then and you risk regretting not buying in when it was cheaper.

    The final piece in the puzzle will be regulatory approval. I think this step will actually be easier than many think. Tesla is arming themselves with a war chest of data, that can prove that their computer-driven cars are statistically safer by an order of magnitude, compared to that of a human.

    For highway driving Autopilot has already surpassed human ability by some margin and once low-to-mid speed driving also achieved that benchmark, it’ll be hard to say no to.

    If a Government official, responsible for safety on our roads, was to deny Tesla from shipping autonomous driving, they would then be saying no to a technology that could cost lives. This would be completely counter to their goals of reducing the road toll, in what could be the first solution to actually drive it to zero.

    All things considered, FSD is not finished right now, with many features still in beta, but I would still recommend you buy it now.

    I can see the day where some cars can drive themselves and others can’t. I’d much rather own a car that offers autonomous driving, as that presents many options not available when you are responsible for operating the vehicle.

    More information at tesla.com

    Jason Cartwright
    Jason Cartwrighthttps://techau.com.au/author/jason/
    Creator of techAU, Jason has spent the dozen+ years covering technology in Australia and around the world. Bringing a background in multimedia and passion for technology to the job, Cartwright delivers detailed product reviews, event coverage and industry news on a daily basis. Disclaimer: Tesla Shareholder from 20/01/2021

    12 COMMENTS

    1. Thanks Jason, great summary. I’m still sitting on the fence and thinking about it, and will obviously to make a final decision within weeks to avoid the rise. I’m very tempted, but usually need to fully justify such a decision , and that’s where I’m struggling a little. Basic autopilot is absolutely fabulous, and I use it nearly all the time, but I’m still very unsure about FSD, especially as I’d then lose it if I for example eventually trade the car on a Model Y for example.
      Anyway, great review, thanks!

      • Thanks very much David, Autopilot is seriously great. Lane centering is a massive step forward compared to most who use lane keeping assist or even worse, just an alert/warning.

        The combination of Summon (slow), City driving (medium) and NoA (fast) should combine nicely to provide an experience that feels very close to full autonomy. Then it’s just down to the edge cases.

        Something I didn’t touch on the article is the length of time you’re going to keep the car. It definitely only makes sense if you’re holding on to it, if you plan on flipping it / upgrading to a Model Y, then I’d wait till then.

    2. Thanks also Jason, excellent overview: Like yourself we have recently splurged on FSD for our Performance minus. One issue that I find more generally involves speed sign recognition. Mostly a confusion around responding to 60 and 80 Kph signs, and also a failure to even notice the higher overhead electronic speed signs on freeways or tollways, Our BMW i3 I find actually does a much better job in regard to speed signs and is something I’d say if Tesla improved could make autopilot use rather more successful. Your thoughts?
      Kind regards.

      • Hi Gary, I’ve only got 1 street on my regular driving that has this issue. I have heard reports from other that they experience it a little more often. Thankfully there is a change coming where Tesla believes they can use computer vision to read speed signs (something the Model S did originally). This should resolve the issue and I hope have it slowing ahead of the sign, whereas now it slows only after it enters the new speed zone, potentially risking a fine.

    3. Many thanks, Jason for your very well-expressed opinions. For some weeks I’ve been evaluating whether to invest in FSD. Your overall descriptions and favourable recommendations have helped convince me to do it now – and before the price hike !

    4. Thanks for the great articel. I,m just wondering if getting the FSD will affect the value of my car (2018 Model X), when Im going to sell it one day. Getting the latest HW might be a plus if you resell your car. It belongs to the car, not the owner, right? I got an offer to upgrade for 2600€ just this week… still thinking…

      • Hi Tim, having FSD should definitely positively impact the price of your car as a car that can partially (and eventually fully) drive itself is definitely functionality that should be accommodated in the price and appreciated by a potential buyer. It obviously won’t make it worth more than a brand new one, so just set your expectations accordingly.

        My advice would be to go for it if you can find the money, as it will only get more expensive from here on out.

    5. It’s not 10+ degrees inclines that smart summon can’t handle. It can handle only up to 5.75 degrees, and anything over that smart summon will fail. You mixed up 10% grade of the incline with degrees.

    6. A mere observation: it’s not a 10+ degree inclines that smart summon can’t handle. Anything over 5.75 degrees, and smart summon will fail. You’ve mixed up 10% grade of the incline with degrees.

    Leave a Reply

    Ads

    Latest posts

    Reviews

    Related articles

    techAU