The xAI API is incredible.
I just created an AI assistant that can fetch news content from URLs and write a post about it on my own writing style.
Super simple to set up and you can try free. Here’s how:
Starship made a controlled reentry, successfully making it through the phases of peak heating and max aerodynamic pressure and demonstrating the ability to control the vehicle using its flaps while descending through the atmosphere at hypersonic speeds
사회적으로 말해보자면 결혼하��� 이게 좋고 저게 좋고 이건 좀 안좋을 수도 있을 것 같아..라고 대답하겠지만, 더 정확한 정보 전달이 우선 목표라면 사실 그것보다는 "우리 좋다, 안좋다의 정의부터 먼저 해보고 이야기해볼까? 너의 좋다, 안좋다는 무슨 의미야?"라고 말해야할 것이다.
The existing factory space was already allocated to vehicle, battery and cell production.
And this is not a matter of tucking a few computers into a corner. You need a kW of power and cooling for each GPU, which would mean 12MW of power and 12MW of cooling.
The south extension is custom-built for heavy power compute and cooling (the cooling towers are huge). Initial system is 50k H100s, plus 20k Tesla HW4 AI computers and massive video storage.
Important to note that FSD training requires massive liquid-cooled Tesla AI compute and video storage, not just Nvidia.
23년 11월 일론머스크,
"차량용 AI 컴퓨터가 모델을 실행할 수 있다면, 테슬라는 아마도 지구상에서 가장 많은 양의 실제 사용 가능한 추론 컴퓨팅을 보유하게 될 것입니다. 미래의 로봇택시에서도 자동차는 주당 약 1/3의 시간만 사용되며, 나머지 2/3는 SETI와 같은 분산 추론에 사용될 것입니다."
$TSLA
FSD 12.4.1 releases today to Tesla employees. If that goes well, then it will be released to a limited number of external customers this weekend.
There are a massive number of changes to this build. It should arguably be called v13, but we’re sticking to 12 😂
Two other versions are in earlier stages of testing: 12.5 and 12.6, which could be called v14 and v15. We are starting to get to the point where, once known bugs are fixed, it will take over a year of driving to get even one intervention.
이것에서 중요한 인사이트를 얻을 수 있음
"Obviously, 5GW of AI training compute is enormous by current standards, but is only about 5% of total Tesla AI compute."
트레이닝 컴퓨터보다는 추론컴퓨터 (HW4) 가 테슬라의 AI 비즈니스의 가장 중요한 자원이 될 것이라는 의미
$TSLA
Training compute for Tesla is relatively small compared to inference compute, as the latter scales linearly with size of fleet.
Perhaps the best way to think about it is in terms of power consumption.
When the Tesla fleet reaches 100M vehicles, peak power consumption of AI hardware in cars will be ~100GW. Training power consumption is probably <5GW. These are very rough guesses.
Obviously, 5GW of AI training compute is enormous by current standards, but is only about 5% of total Tesla AI compute.
There is a path for Dojo to exceed Nvidia. It is a long shot, as I’ve said before, but success is one of the possible outcomes.