@Tesla Summon ran through the yellow tape. Hi @Tesla and @elonmusk, I post this because I want Tesla to do a better job in developing #FSD. I love the product and use FSD regularly. In the post, you see me use it 88% of my daily drive. The Summon feature is a factor that made me buy Tesla also. FSD is good, but Summon still needs to improve. This time it ran through the tape; another time it stopped right in busy parking. I tagged the videos so you can look into it and improve it. Thanks with respect.
@elonmusk Thank to you! I take trip for three hours without touching the wheel and pedal . It’s 154 miles relaxing helping us to have a good finally road trip. We can exchange conversation with kids with a free mind set.
Born without hands.
25,000+ miles… most of them on FSD.
@real86hands just shared how Tesla Full Self-Driving has completely changed his life.
One button.
Hands free.
Smile on his face every single time he gets in the car.
This is what the technology is actually for.
Tesla has released another new video of FSD (Supervised) avoiding accidents/preventing pedestrians from getting across multiple European countries, including many where FSD hasn’t been approved yet.
Delaying FSD’s approval will cost lives.
Tesla $TSLA delivered 480,126 vehicles in Q2, well above estimates around 396K to 406K.
Q2 breakdown:
Produced: 451,758 vehicles
Delivered: 480,126 vehicles
Model 3/Y deliveries: 467,762
Other model deliveries: 12,364
Energy storage deployed: 13.5 GWh
AI5 (Tesla's new inference chip for FSD, Optimus & Robotaxi) just taped out—optimized for real-world edge AI with full HW/SW co-design.
Vs top chips (NVIDIA Hopper H100, Blackwell B200, AMD MI300X, Google TPU v6, Intel Gaudi3):
- Single AI5 ≈ H100 inference perf (~2k-2.5k TOPS effective) at ~150-250W vs 700W+.
- Dual AI5 ≈ Blackwell-level.
- 10x better perf/$, 3-5x efficiency.
- Costs peanuts, fits in a car (not datacenter-only).
Massive leap from AI4 (40x in key metrics). Production ramps H2 2026.