[Tesla] Add 'Find My Friends' feature to in-car map

@sdamico Do we include Tesla and Rivian in this? Some can't live without CarPlay and others swear by TeslaOS. Fundamentally car manufacturers and Big Tech are in a fight for platform control, not good products - Meta learned this lesson HARD. Conquer and figure out what to do on the land later. My biggest Tesla/Rivian criticism is that (following in Steve Jobs' initial footsteps), they don't allow and foster third party development. The TAM for good apps in a car is massive. Think about the vendor lock in effect if, say, Tesla had a "free parking finder" app, or a "Find My Friends on Tesla map" feature. Find My and iMessage work well on iPhone because they have critical same-manufacturer mass. Tesla is getting there. Apple and Google CarPlay are attempts to plug this gap, but I see CarPlay and Android Auto as attempts to steal market share from the OEMs and create phone lock in, as opposed to unlocking in-car value. For one, large tech companies are WAY too risk averse to build an effective in-car system. The onerous approval process + functional restrictions (No custom UI elements!!!) go FAR beyond the FMVSS "no video or television visible to the driver" restriction. The OEMs' model of an in-car experience is the one where they show sponsored content from "in-network" bundleware providers. Apple/Google only had to deliver enough value to stay. on the user's screen, and get the user to demand the feature. Having Music and Waze was good enough for this. While Apple milked CarPlay for MFi license fees, Google decided to close source theirs too, unlike Android core (this generally corresponds to the time when Google dropped their "don't be evil" motto). The closed ecosystem made it exceptionally difficult to create integrations - and I do not blame car manufacturers for the terrible integrations. The fundamental API surface around CarPlay is "Stream this H264 video over TCP over USB, and stream button press events back". You cannot build a good integration with a model like that. There's a lot of similarity between the American auto OEM software model and the pre-iOS carrier-feature phone model. Under the feature phone model, you 1) predict, or dictate, what features the users want 2) Convince the user to buy the car/phone based on advertising of those features (exclusivity, etc). Under the iOS model though, it's less about giving the user a tool and asking them to find a use, and more about asking the user what the use is, and letting them discover a tool (often made by someone else - "there's an App for that") to address it. This model can be unfamiliar to Auto OEMs because the execs and engineers tend to be on the older (sorry generalization) side. From experience, their average mental model of computers is more likely to be "discrete features/paths", i.e. Click start, file explorer, go to C, get a file vs "files are organized by folders, go find one that's likely got what you need in it." This is going to inform the way the software is created, where specific use cases are unintentionally fossilized in the code. Furthermore, they're just more familiar with the "do market research, partner with others, ship" model which is popular in hardware. Most don't explore technology on their own - feature quality is evaluated by checkbox (exec understands the value of the feature but does not personally need it; thus does not value it being fast, rather values the feeling and effect that it is possible) or without comparison (voice assistants from OEMs, like Mercedes Me). A lot of OEMs mistakenly believe that their features are selling points (or that they can make money by upselling heated seats DLC), instead of letting others build their platform and making their money by swaying customers. Another issue in embedded software is engineer vs user driven design. Apple and BigTech use the user-driven model; start with HCI people and UI/UX designers (artists) to capture the user's emotions and needs, then break it down to functionality and implementation, and give the engineers a rigid blueprint for the end product. Embedded tends to use the engineer driven model; define technical/feature requirements, build the system as if it were an analog fluid power system, then subdivide functionality, map the components to UI, and then build. UI/UX is often outsourced, and is applied as a theme over the finished software. Thus, it's not possible to structurally build better integrated or more logically organized experiences; the "chassis" is fixed and only the skin can change. So on the issue of Chinese OEMs, roughly there are a few strategies. Tesla is unique in that they don't really have installable apps - every app is part of the base OS image they ship to all of the cars. Companies like LiXiang have some installable apps (customized specifically for them by the app makers, mostly video streaming etc.) It's likely that it's a similar situation to HTML5 based Smart TV apps. On the other hand, Xiaomi HyperOS, Huawei HarmonyOS etc have native phone app ecosystems, which allows them to run the same apps (to a degree). Furthermore, almost all of the systems support some form of screen mirroring (not necessarily CarPlay, straight mirroring). China lacks the American FMVSS prohibition on television - I'm guessing this is why they can implement mirroring while Apple has to lock down CarPlay to prevent apps that put it in non-compliance (of course, Tesla's solution of providing a browser where the video tag won't play is a much better tradeoff profile!, and yes you can work around this by displaying image sequences in Canvas, which is what Tesla Android does to stream a RasPi to the screen.) Chinese OEM software tends to copy the Apple user-driven design approach, but in a roundabout way: by copying Western design trends. I would put the usability of their software higher than American OEMs (bad development and design), but below Tesla, who has a consistent vision for how the system should look and feel. There is a vast pool of talent in China, so implementation of features is rarely the bottleneck. However, there is the same lack of strategy. I wouldn't necessarily say they are not surpassable by a sufficiently dedicated software team in the West. One major roadblock is that in China, if you go down the "partner apps" route, many more apps are willing to develop a custom integration for you than in the West, where popular apps don't even have iPad versions. My personal vision for how automotive and appliance OEMs can do well in the market (subject to change): 1) Build a software team isolated from corporate processes. Build a company within a company, and hire people familiar with the Apple/BigTech style of development. Start with UX/UI people, HCI people, do lots of testing. Bring on evangelists and power users who can tell you when something looks right or wrong, and really define your brand's UI strategy and tone. Build strong values to guide your software. This is how Apple always wins the "just works/feels nice" arguments. 2) Offer CarPlay and Android Auto. This one's an easy slam dunk to the customers who walk away from your product because it doesn't have this. 3) Especially for smaller products, have some kind of an ecosystem that developers, power users, and evangelists can join and add value for your customers. Early adopters, while not representative of the average user, create the critical mass of content that platforms need to grow. Emerging markets can be opportunities here. This also serves as an escape hatch for "missing" features! Tesla's custom browser is a very good example of a technical solution to certification issues, and WeChat mini programs are a very good example of how apps create lock-in, increase time in-app, and provide easy-to-develop experiences based on familiar Web technologies. 4) Try very hard to abandon the urge to do short term plays like upselling, partnered/sponsored content, etc. to build trust. A good software platform is like a garden. Not a short term stock value boost, a long term play to keep users coming back to the product.