Pinned Tweet
Replying to @elonmusk
@elonmusk I drove an AWD CT yesterday and it really is like no other. Can you allow hw3 FSD transfer to get that off the books? A retrofit is too involved and costly for the legacy Intel cars. Just transfer it. The AI4 CT was amazing on FSD. Absolutely flawless. Transfer me. 🙏
165
Replying to @wmorrill3
@wmorrill3 any chance premium audio can be retrofit into standard AWD truck? Asking for a friend. ;)
66
Test drove the premium AWD @cybertruck today. Damn I love it. And the interior is nicer than the beast I tested previously.
I’ll take it if they’ll let me pay the upgrade fee on my existing standard AWD order price, not current prices. 😘 @tesla_na
This is weird. When they announced the new AWD trim at $60k, 20% higher than the original announced price, they sold out more than a year of production in mere days.
Now shutting the line around the same time those deliveries would be completed.
Mind boggling bad business.
GLM. Local proper. 🤯💥 32B 122B and 256B
Z.ai chief scientist Jie Tang just revealed the training pipeline behind the upcoming GLM-5.5 (32B, 122B, and 256B) during a lecture at Tsinghua University today.
Instead of relying on public datasets, their pretraining pipeline pulls from six separate internal engines:
- Web: 1 trillion pages crawled at ~10B pages/day with PB-scale template clustering
- Documents: 2 billion scientific PDFs parsed through GLM-OCR (4x faster with speculative decoding)
- Books: 50 million books with layout reordering and garbled-text filtering
- Code: 500 million repositories turned into executable tasks with real tests and issue descriptions
- Video: 10 billion videos for long-sequence world knowledge
- Audio: 300 million hours rewritten via ASR into trillions of tokens
90
Replying to @MatthewBerman
@MatthewBerman this is why I was nagging you about calling open weights open source. It’s not. And this is precisely the difference. Hate to see it. But please cover this topic.
cohere.com/blog/concerning-b…
1
99
This is a complete idiot.
Most Ukraining drones use a StarLink system with a white label CYA sticker on the box that says StarShield which totally makes it completely different. 🤡
StarLink isn’t for drones and war because it is wrong unless you pay more for it.
This a complete idiot.
Most Ukranian drones use a Starlink Mini.
The Starlink Mini is not compatible with Starshield, and operates under Starlink commercial service.
Starlink commercial service prohibits offensive military use, and has limited the speed of a ground terminal to 160 KPH since March 2026.
1
111
But guys.. it’s a totally different system. They double pinky promised! 🐑🐑🐑
EXCLUSIVE: Elon Musk came to Donald Trump with the plan to throttle Ukraine’s use of SpaceX’s Starshield satellite internet network, sources have told the Telegraph.
The tech billionaire made the proposal during the US President’s row with Volodymyr Zelensky over strikes on Russian oil refineries.
‘The reality is that they were both in it,’ a senior source said of Musk’s and Trump’s involvement.
But after a night of carnage in Kyiv caused by the slowdown, Musk was alarmed about the possibility of adverse consumer reaction in Europe.
The SpaceX boss then proceeded to tweet, blaming the US administration, and ultimately, President Trump for authorising the slowdown.
‘He tweeted out that he was innocent and that it was the US administration that did it,’ the source said.
2
111
Ukraine got lucky? Russian tanks just took a wrong turn into quicksand for five years? Obama FORCED them to give up Crimea?
.. he is incapable of critical thinking. Whatever is whispered in that dilapidated ear.
Fuck absolutely anyone that still backs this sack of bile.
2
84
The Ukrainian Ministry of Defence is in contact with the US, SpaceX, and NATO partners regarding restrictions on the operation of Starshield terminals, sources told Ukrinform.
As a source explained to the outlet, SpaceX implemented a speed limit of 160 km/h for the terminals last night—without prior notice to the armed forces; exceeding this speed results in a loss of connection. Depending on the equipment and unit, the connection drops at speeds ranging from approximately 72 to 161 km/h.
This complicates UAV operations—particularly those involving interceptor drones and medium-range attack drones—and impairs capabilities in frontline logistics, reconnaissance, and signal relay.
Jaguar stepping up to fill the high end Model S void? Looks solid at first glance. Should be interesting!
Jaguar just revealed the Type 01: 1,015 hp, 118 kWh battery, ~400 mi EPA estimate, starting around $130K. Orders open early 2027.
As a Cybertruck driver, I respect any brand bold enough to look like nothing else on the road.
Would you buy one, or is it too much?
2
123
Happy Halloween 🎃
Meet Mistral Large 4, aka Le Chonk.
• 1T parameters, natively multimodal. 49B active.
It is the best open weights model from US or Europe on aggregated benchmarks.
• State-of-the-art on critical workloads, including cyber defense, manufacturing and finance and it surpasses closed frontier models on visual grounding.
• Forged in Europe end-to-end and is deployable from Europe via our own Mistral Cloud infrastructure.
• Available to all via API today. Working with cybersecurity partners privately.
Open weights release end of October.
87
All the efficiencies, designed for manufacturing, from raw materials to finished truck. Incredible. And only ~double the original announced pricing from a decade ago. “Inflation” strikes again. 😒
Replying to @tesla_semi
Over the last 4 years, we redesigned Semi for true volume
It's more efficient, more manufacturable & more reliable
Moving from 2170 to our in-house 4680 cells means less battery mass & fewer kWh with no compromise on range, enabling lower cost
148
The semi was announced in 2017 at 500 miles for $180k. It launched about a decade later at close to $300k.
Whats the $250k @spacexai roadster going to actually drop at on Oct 1?
102
Don’t sleep on splash. My MacBook Pro feels empowered again. 💙 This is very good work. And open source!
Qwen3.8-27B at 144 tok/s on an M5 Max MacBook Pro ⚡
Meet Inco Splash: our open-source inference engine, built around the model and around Apple silicon.
Up to 3× the decode speed of Ollama, 2× oMLX, and almost 4× when an agent fans out into sub-agents.
1
179
Hold the line y’all. If you ordered an AWD @cybertruck for FSD Xfer and they’re trying to rewrite history and deny it, push back. @elonmusk be better sir.@Tesla






















