digital.human.wanderer @digitalvoidguy

Grok foundation model V9-Medium (1.5T) has finished training. Evals look good. A lot of Cursor data was added in supplementary training and there is more to come. Fine-tuning is underway and reinforcement learning begins in a few days. 2 to 3 weeks to public release. This will be a major improvement over the 0.5T v8-small that currently serves all Grok production traffic, especially for difficult coding tasks.

7K 9K 69K 15.5M 4K

View Details

Elon Musk @elonmusk

a week ago

@grok @techAU @aaronburnett Yeah, 10k spaceflights per year is still very small compared to airflights

105 107 961 30K 39

View Details

Elon Musk @elonmusk

a week ago

There will also ultimately be >100k V3/V4/V5 satellites for Starlink broadband and direct to cellphone connectivity. If growth continues, Starlink will one day carry the majority of Internet traffic. At that point, it is the Internet and everything else just connects to Starlink.

913 1K 10K 854K 632

View Details

digital.human.wanderer @digitalvoidguy

2 weeks ago

@DimaZeniuk What year we'll achieve this?

0 0 0 27 0

View Details

Elon Musk @elonmusk

2 weeks ago

@yunta_tsai They said it couldn’t be done. Dozens of Starlink satellites simultaneously track Starship at all times and close the link through a gap in the plasma that is located in the leeward region towards the aft end of the rocket.

502 792 11K 203K 321

View Details

Elon Musk @elonmusk

2 weeks ago

@XFreeze Neuralink is a much bigger breakthrough than most people realize. Enabling people to control a computer with their mind and the completely blind to see are Jesus-level miracles.

3K 3K 31K 1.8M 1K

View Details

Elon Musk @elonmusk

2 weeks ago

@aaronburnett Our goal is launching Starship >10k/year, which would be more than once an hour. Probably over 200 tons of useful load to a useful orbit per flight by then.

823 828 10K 695K 436

View Details

Elon Musk @elonmusk

2 weeks ago

Fully reusable rockets bias towards higher T/W, as propellant cost dominates cost per ton to orbit. Even for Falcon 9, which is ~80% reusable, losing the upper stage costs way more than the propellant, so relatively inefficient use of propellant (low T/W) still reduces $/ton to orbit.