3 ms·

I am a Tesla owner, and yes, I regret the purchase of FSD. I did and still do consider it a bit of an investment in a future I'd like to see, and I'm quite happ

I am a Tesla owner, and yes, I regret the purchase of FSD. I did and still do consider it a bit of an investment in a future I'd like to see, and I'm quite happy I also _actually invested_ in Tesla when I bought my model 3, but over the last 4 years I've learned something that has changed my mind considerably: AI does not step forward progressively. Normally, engineers spend ages improving things, testing, improving again. With AI, it's far more of a "let's just re-train this black-box on a totally different set of data, wind up with a totally different black-box, and see if its better!" The difference in driving quality and intelligence between even minor version updates is dramatic. I very intentionally _do not_ update my Tesla until the release is ~6 months old. This... sucks. I am still very bullish on Tesla and EVs, and on technology in general - unlike many I am not full of hate that this didn't work out, nor do I particularly think FSD is _dramatically_ more dangerous behind the wheel than say my grand father was before we convinced him to stop driving, but I do think I have more or less zero patience for AI engineers these days. The failure by Google and others to deliver even a moderately usable voice assistant is all the proof I need that FSD is _years_ away.

mort96 5y ago
Yeah, I think most people who haven't worked with machine learning may not realize this, and it's extremely important. It can be pretty easy to reach 90%, but impossible to reach 95%. People tend to assume some sort of progression, where you just "iron out" the remaining quirks, but that only applies to traditional engineering. You can gather 100x the training data and build a model with 100x the number of nodes and get a 10% average improvement with some cases getting worse and others getting better. I think the current state of FSD is extremely impressive. There are clearly some extremely talented engineers working on that system. But my guess would be that it simply can't be incrementally improved until it's road-safe, in much the same way that GPT-3 is insanely impressive but can't be incrementally improved until it can write a whole human-quality novel. I suspect a new wave of radical rethinking of machine learning might be necessary, and that might be decades away.

mort96 5y ago

Yeah, I think most people who haven't worked with machine learning may not realize this, and it's extremely important. It can be pretty easy to reach 90%, but impossible to reach 95%. People tend to assume some sort of progression, where you just "iron out" the remaining quirks, but that only applies to traditional engineering. You can gather 100x the training data and build a model with 100x the number of nodes and get a 10% average improvement with some cases getting worse and others getting better. I think the current state of FSD is extremely impressive. There are clearly some extremely talented engineers working on that system. But my guess would be that it simply can't be incrementally improved until it's road-safe, in much the same way that GPT-3 is insanely impressive but can't be incrementally improved until it can write a whole human-quality novel. I suspect a new wave of radical rethinking of machine learning might be necessary, and that might be decades away.