Posts
SortierungNeueste zuerst
Zufälliger Post
Wall Street Mav
@WallStreetMav · 29. Juni 2026
Feminism was created to dismantle the family, not to liberate women.
Elon Musk
@elonmusk · 29. Juni 2026
Feminism and lack of children has killed the happiness of so many women
3077
5903
46.620
1,4 Mio.
Whole Mars Catalog
@wholemars · 29. Juni 2026
they’re still going to overtake Tesla this year though, right?
Sawyer Merritt@SawyerMerritt· 29. Juni 2026VW plans to end its self-driving partnership with auto supplier Bosch as part of a wider push to cut costs and after VW determined the technology was not competitive with other self-driving systems such as @Tesla's FSD.
This follows a report that VW plans to cut 100,000 jobs
Elon Musk
@elonmusk · 29. Juni 2026
Guess not
167
117
3227
94.734
Visegrád 24
@visegrad24 · 29. Juni 2026
🇩🇪German feminist Verena Brunschweiger promotes the slogan "My lineage ends with me."
She calls on White people not to have children, claiming that Western pro-natalists only want to "control women, and keep refugees out."
Brunschweiger says Europeans are responsible for poor
Elon Musk
@elonmusk · 29. Juni 2026
Deal with traitors first, then invaders
2165
8355
70.416
1,2 Mio.
Xristine
@noen25083 · 29. Juni 2026
Back then we thought of C as being wasteful compared to assembly code.
Interesting that C/C++ now is the fast code to get back to.
Elon Musk
@elonmusk · 29. Juni 2026
😂💯
We are using some assembly (and direct ASIC primitives) for operations that are done trillions of times per second
65
64
731
66.152
stevenmarkryan
@stevenmarkryan · 29. Juni 2026
The more I use Grok and Grok Build, the more I think about (and am pained by) how insanely LOW BANDWIDTH human communication is.
Really wish could dump what's in my brain about 1000x faster.
I've switched to voice prompting for speed but it's still way too slow.
(Neuralink has
Elon Musk
@elonmusk · 29. Juni 2026
Neuralink is needed to solve the super low & lossy output bandwidth of humans.
Our input bandwidth, thanks to vision, is many orders of magnitude higher than our output bandwidth, but could also be greatly improved.
By addressing these bandwidth limitations and greatly increasing our effective cognitive capacity, human-AI symbiosis will be far better, improving the probability of a great outcome for humanity in the future.
998
1428
7373
1 Mio.
Arthur MacWaters
@ArthurMacwaters · 29. Juni 2026
The reason engineering and business are resistant to Marxism is that they require engaging with the real world
The Rabbit Hole@TheRabbitHole· 29. Juni 2026Faculty at American universities:
— 7.7% Marxist
— 25.8% Socialist
Elon Musk
@elonmusk · 29. Juni 2026
Exactly
100
61
912
68.875
Whole Mars Catalog
@wholemars · 29. Juni 2026
As the model gets better at writing code, the usefulness of abstraction layers designed to make things easier for humans trends to 0
I would probably be concerned with memory safety, but the models are already very good at identifying and fixing buffer overflows etc
Elon Musk@elonmusk· 29. Juni 2026@aaronburnett Truly massive gains will come in ~3 months when the entire training and inference stack is written in C/C++ and massively simplified (most software layers will be deleted completely) and we exact-map Grok to work incredibly well on a GB300
Elon Musk
@elonmusk · 29. Juni 2026
With great power (direct memory management) comes great responsibility
224
138
2240
125.474
Harsh But Mostly Thoughtful John McGrath
@HarshThoughtful · 29. Juni 2026
d this.
We wrote Telescript teleprompter software specifically in C and C+ to avoid the layers and “speak” directly to the firmware.
It was key in the early computer/video days (early 1990s) to have zero delays when important people depended on the words they had to say in
Elon Musk
@elonmusk · 29. Juni 2026
Same. I worked video games in the 90s with no graphics accelerators. Every CPU clock cycle mattered.
79
61
1020
80.980
Ryan Wang 🇹🇼
@ryanwang · 29. Juni 2026
xAI is undertaking a major overhaul of Grok’s training and inference software stack. The core of this transformation is rewriting the entire system from high-level frameworks into C/C++, drastically eliminating intermediate abstraction layers, and performing exact mapping
Elon Musk@elonmusk· 29. Juni 2026@aaronburnett Truly massive gains will come in ~3 months when the entire training and inference stack is written in C/C++ and massively simplified (most software layers will be deleted completely) and we exact-map Grok to work incredibly well on a GB300
Elon Musk
@elonmusk · 29. Juni 2026
In some cases, where we don’t have access to the source code, eg network switch software, we are also decompiling and modifying the binaries to improve performance
419
858
4805
714.923
Around Town Memes
@chuckdpete · 28. Juni 2026
Elon Musk
@elonmusk · 29. Juni 2026
🎯
344
343
11.416
323.655
Hunter Ash
@ArtemisConsort · 28. Juni 2026
Projections 74 years into the future are a joke. The world will be completely unrecognizable much sooner than that. No one knows what’s going to happen.
Leading Report@LeadingReport· 28. Juni 2026As President Trump expands immigration enforcement policies, Census Bureau projections show that sustained extremely low or zero net immigration could leave the U.S. population more than 107 million smaller by the end of the century, per NPR.
Elon Musk
@elonmusk · 29. Juni 2026
💯
158
81
1403
93.617
Elon Musk
@elonmusk · 29. Juni 2026
Yes
Elon Musk
@elonmusk · 29. Juni 2026
@scaling01 Every aspect of the 2T model is improved over the 1.5T, so it will be much more than just a 1/3 increase in params
166
162
2651
149.979
Lisan al Gaib
@scaling01 · 29. Juni 2026
Sounds like we will get the 1.5T model in July and the 2T model in August
Elon Musk@elonmusk· 28. Juni 2026@beffjezos In addition to their excellent and unique training data, the Cursor team is also making major engineering contributions to v9 SFT & RL. It’s an honor and a pleasure to work with them.
For this 1.5T run, Cursor data was added in supplemental training, which is not quite as good
Elon Musk
@elonmusk · 29. Juni 2026
Yes
94
105
2274
193.431
Ron DeSantis
@RonDeSantis · 29. Juni 2026
Ivy League universities are hedge funds with indoctrination camps as a side hustle.
End Wokeness@EndWokeness· 29. Juni 2026Harvard U: $56.9 billion in endowments
Princeton: $36 billion in endowments
UPenn: $24.8 billion in endowments
Cornell: $12 billion in endowments
Brown: $8 billion in endowments
Yale: $44 billion in endowments
These are the same people demanding Elon "end world hunger" (lol)
Elon Musk
@elonmusk · 29. Juni 2026
Yup
187
163
3386
222.578
Guillermo Rauch
@rauchg · 28. Juni 2026
You don't need a LinkedIn, you need a page on your website describing and linking to what you shipped
Elon Musk
@elonmusk · 29. Juni 2026
💯
190
92
1469
142.832
Blake Scholl 🛫
@bscholl · 28. Juni 2026
Build the future you want to live in.
Elon Musk
@elonmusk · 29. Juni 2026
Yes
249
91
1462
99.074
Aaron Burnett
@aaronburnett · 28. Juni 2026
Let’s see what calling in the aerospace/rocket engineers to work on AI models does for Grok.
Elon Musk@elonmusk· 28. Juni 2026@beffjezos To be clear, I’m not saying the Grok v9 foundation model will be mind-blowingly better than anything, but it will be a solid workhorse in the same league as Opus.
And the SpaceXAI cadence of model and harness improvement is speeding up tremendously, particularly due to a few
Elon Musk
@elonmusk · 29. Juni 2026
Truly massive gains will come in ~3 months when the entire training and inference stack is written in C/C++ and massively simplified (most software layers will be deleted completely) and we exact-map Grok to work incredibly well on a GB300
2376
1846
12.027
1,6 Mio.
Peter H. Diamandis, MD
@PeterDiamandis · 28. Juni 2026
We're sending data centers to space because we live in a planet that cannot generate enough power fast enough to satiate the need.
Orbit can solve what Earth cannot.
Elon Musk
@elonmusk · 29. Juni 2026
Space is inevitable if you wish to ascend to a K2 civilization
426
255
3415
193.028
Aaron Burnett
@aaronburnett · 28. Juni 2026
Our estimates for Starlink capacity/revenue are among the most bullish in the industry. The more I dive into the AI compute side the more I’m realizing we likely underestimated Cursor/grok.
Shay Boloor@StockSavvyShay· 28. Juni 2026$SPCX is expected to approach ~$100B of revenue by 2028 which would make SpaceX one of the fastest-scaling businesses in the world at this size.
The reason that ramp is even possible is that SpaceX controls the launch layer, uses that launch capacity to scale Starlink, uses
Elon Musk
@elonmusk · 29. Juni 2026
I would be disappointed if SpaceX did not significantly exceed these milestones
339
314
4830
1,3 Mio.
Elon Musk
@elonmusk · 28. Juni 2026
True
House Republicans@HouseGOP· 28. Juni 2026100 million dead.
That’s the verdict on Marxism. History already ran this experiment.
Never again.
514
631
3685
669.942