@alxfazio i was thinking the same to switch from windtre. it doesn’t necessarily mean the deal is over but i remembered this was a thing and it’s a funny connection haha. we’ll see in the coming days i don’t think italy will be left behind (hopefully at least)
The algorithm already uses a Grok-based Transformer for the Phoenix retriever while gathering and presenting out-of-network posts and this can be customized with the newly introduced custom timelines feature. Using a connector that is custom to each user inside the custom timelines page can be implemented rather very easily without modifying the algorithm.
@aarniodev@levelsio@xai You guys need to chill :) It's already updated!
High reasoning effort is awesome. Any plans to bring the reasoning_effort picker to 4.3 on Grok Web too? On high effort, it just solved a problem none of the previous Grok models have ever solved before. Please consider this.
@tesla_na@SawyerMerritt@cybertruck This feature should roll out to 2020–2024 Model Ys worldwide as well, as there is no hardware limitation compared to the implementation on the refresh Model Y
@saitamasma@elonmusk@LiveSquawk This is not the same and is a very dumb take. Apple supports devices in the thousand k range for over 7 years, as Tesla is doing the same. But it does not sell software for 8k and does not provide the software at all. FSD markets outside NA don’t even receive FSD 12.6.4 for HW3
@MdeZegher@TeslaCharging Could you elaborate on what the “units” are, Max? Are you referring to the AI4 inference on the vehicles which will be active during Supercharging sessions, or will these be external units positioned on Supercharger sites?
@Sentient_Atoms@elonmusk@LiveSquawk In the European market, the AI4 switch for the Giga Berlin Model Y happened in April 2024, meaning cars as young as 2 years and 11 months old are equipped with HW3