That’s an interesting thought. I think it depends on moral dilemma answer.
Better to save people no matter what they drive or better that EVs are pushed in every way possible so there’s people actually left to save?
Well that's quite a way to frame it! You're pushing me to think a little deeper. Thanks.
A few thoughts...
1 Roll out is likely to be model by model. I would expect that FSD would need to be custom calibrated to each model so that it can safely drive that model.
2 Tesla does does not have the compute and staff to adapt FSD to all models at once.
3 if 1 and 2 are correct, then there is a practical need to prioritize some models over others.
4 Well-calibrated FSD improve safety for all, not just the occupants of the FSD vehicle but occupants of other vehicles and pedestrians as well. So the ethics of improving road safety argue for a rapid and well-calibrated deployment in the interest of all who share the road.
5 FSV has been developed and optimized for EVs. Adapting FSD to ICE vehicles may be suboptimal. Arriving at a certain safety threshold for an ICE model may require substantially more cost, compute and engineering talent that porting FSD to another BEV.
6 If 5 is so, the friction of porting to ICE could slow the scale up desirable in proposition 4. Slowed deployment reduces safety benefits for all on the road including pedestrians and occupants of non-FSD vehicles.
7 The cost of porting FSD to a model is likely a substantial investment on the part of the OEM implying multi-year payback. From the perspective of an OEM, does it make sense to make this investment on an ICE model which is at risk of obsolescence within a few years? Probably not. However, some OEMs may entertain the delusion that they will have robust demand for a given ICE model that will last for more than a decade, if only they can secure autonomy tech. Tesla likely does not share that vision of ICE sales persisting out to 2035 or so. This misalignment of time horizons might not be healthy for partnership, creating other frictions that slow the scale up of FSD.
8 Even if Tesla and an OEM are unconcerned about impending ICE obsolescence, deploying FSD to an ICE will likely increase the sales and longevity of that ICE model. This works contrary solving pollution and climate change problems.
9 My basic conclusion is that FSD should be deployed to BEV models first, with priority over ICE. This prioritization seems to make better use of the limited resources for scale up (total road safety), while hastening the obsolescence of ICE (combating pollution and climate change).
By the time autonomy is achieved for all BEV models, will there even be much of an ICE market left?