The US is falling way behind in electric vehicles. If BYD could sell in the US, the US auto industry would be crushed.[1]

What went wrong is that 1) Tesla never made a low-end vehicle, despite announcements, and 2) all the other US manufacturers treated electric as a premium product, resulting in the overpowered electric Hummer 2 and F-150 pickups with high price tags. The only US electric vehicle with comparable prices in electric and gasoline versions is the Ford Transit.

BYD says that their strategy for now is to dominate in every country that does not have its own auto industry. Worry about the left-behind countries later.

BYD did it by 1) getting lithium-iron batteries to be cheaper, safer, and faster-charging, although heavier than lithium-ion, 2) integrating rear wheels, differential, axle, and motor into an "e-axle" unit that's the entire mechanical part of the power train, and 3) building really big auto plants in China.

Next step is to get solid state batteries into volume production, and build a new factory bigger than San Francisco.

[1] https://en.wikipedia.org/wiki/List_of_BYD_Auto_vehicles

I decided to check this out after seeing the discussion here. I had previously misunderstood that it required a Claude.ai plan, but it actually just uses your API keys.

I did a comparison between Claude Code and Aider (my normal go-to): I asked it to do clone a minor feature in my existing app with some minor modifications (specifically, a new global keyboard shortcut in a Swift app).

Claude Code spent about 60 seconds and $0.73 to search the code base and make a +51 line diff. After it finished, I was quite impressed by its results; it did exactly the correct set of changes I would have done.

Now, this is a higher level of task than I would normally give to Aider (because I didn't provide any file names, and it requires changing multiple files), so I was not surprised that Aider completely missed the files it needed to modify to start (asking me to add 1 correct file and 2 incorrect files). I did a second attempt after manually adding the correct files. After doing this, it produced an equivalent diff to Claude Code. Aider did this in 1 LLM prompt, or about 15 seconds, with a total cost of $0.07, about 10% of the Claude Code cost.

Overall, it seems clear that the higher level of autonomy carries a higher cost with it. My project here was 7k SLOC; I would worry about ballooning costs on much larger projects.

Based on the location and focal mechanism of the earthquake (https://earthquake.usgs.gov/earthquakes/eventpage/nc75095651...), this is a strike-slip earthquake on the plate boundary between the Pacific and Gorda/Juan de Fuca plates. Strike-slip earthquakes occur when two plates slide beside each other during an earthquake, usually along a steeply-dipping if not vertical fault. These kinds of earthquakes almost never produce damaging (or even really noticeable) tsunamis because there is no real displacement of sea water by seafloor movement, unlike a thrust or subduction zone earthquake.

The USGS's automated systems calculate the location and focal mechanism/moment tensor pretty much instantly from the seismic network. The system should know that a significant tsunami is unlikely based on the parameters of the earthquake. On the one hand, it's good to be cautious, but on the other hand, a system designed to cry wolf is also self-undermining. Maybe they should have a tiered warning system?