2. Multi-Step Reasoning
— God of Prompt (@godofprompt) 27 mai 2025
Prompt: Explain Tesla FSD’s RL architecture and find edge-case safety flaws (check full prompt in video)
Result:
Gemini: 2
DeepSeek: 0
→ Gemini delivered better logic and depth. DeepSeek R1 fell short. pic.twitter.com/ZpH4gdMxvS
2. Multi-Step Reasoning Prompt: Explain Tesla FSD’s RL architecture and find edge-case safety flaws (check full prompt in video) Result: Gemini: 2
DeepSeek: 0 → Gemini delivered better logic and depth. DeepSeek R1 fell short.


