Year 2012″Electric cars are just expensive toys, they can’t make ‘real’ cars like internal combustion cars.”

1
Year 2012
“Electric cars are just expensive toys, they can’t make ‘real’ cars like internal combustion cars.” → Tesla Model S Success Makes Faster Cars
The year of 2018
“Tesla will never succeed in mass production of Model 3.” → Model 3 successfully mass-produced
The year of 2020
“Electric vehicles will never make a profit even if they succeed in mass production.” → Tesla achieves annual surplus
2023
“Electric cars are only for some early adopters, but it’s hard to enter the mainstream market.” → Model Y becomes the best-selling model in the world (including internal vehicles, literally the world’s best-selling car)
2024
“Camera alone is impossible to drive autonomously.” → Vision-only end-to-end AI method changes FSD and does well
First half of 2025
“Tesla will never be able to launch a robotaxi as planned.” → Launched as planned for a robotaxi service in Austin, Texas
Second half of 2025
“What’s the point of a robotaxi driving well? There are safety personnel.” → Safety personnel will be removed next month
“Robotaxi scale-up (large scale expansion) is not possible.” → It has reached this point.

2
Just a few years ago, there was a meme that no matter how advanced autonomous driving is, you won’t be able to drive in Busan. Looking at the FSD video released today, I drove comfortably through the Yeonsan intersection in Busan. It was a perfect drive that did not require any intervention on the road throughout the entire Busan schedule. Videos close to testimony have already flooded overseas for a few months, but it has proven that there are no technical obstacles to Korean application.
This is the current version of FSD, and v14.3 will be released next month. It has increased parameters by nearly 10 times compared to the previous version, and is expected to be an innovative update that is close to completion.
Double the sensor processing speed improves almost complete driving stability and responsiveness. Introducing real-time inference-based parking functions, solving sudden breaks and hesitation problems at intersections, construction sections, and unpaved roads, etc.

3
Of course, it’s not the end of v14.3. There’s a major update every half year or so, and improvements are being added a lot each time.
A few days ago, a safety guard was caught dozing off in a robotaxi, and technically, there is no reason why it is impossible to remove the safety guard. Now, finally, with Finaly and Real, the scale-up of the robotaxi will gradually proceed in the near future. lol
Even if there’s something I’m worried about, will it be like that a year later?