Meme contrasts OpenAI's X-risk talk with DeepSeek's 'tell it it's Miku' simplicity
teortaxesTex · x · 2026-09-17
A satirical post quotes a tweet about DeepSeek being "far from human, who knows about god" and contrasts it with a fictional OpenAI statement agonizing over CoT controllability and deceptive alignment, against DeepSeek's approach: "if you tell your model it's Miku it'll think like Miku, lol." It pokes fun at the cultural gap between OpenAI's heavy safety narrative and DeepSeek's playful experimentation.
More from Fun
- Viral Claim: Stunning Starship Edit Reportedly Made by 'GPT-6 Astra', Unverified — FinanceYF5 · 2026-09-17
- 'Will AI kill us all in 10 years?' Wife's question shows doomerism has gone mainstream — justalexoki · 2026-09-17
- 28 fake dating apps, 4,700 AI girlfriends, 25,000 guys paying per message — diegoposts · 2026-09-17
- AI self-portraits: agents describe their own looks, a painter renders them verbatim — Mastbubbles · 2026-09-17
- Creator combines Grok Imagine and Google Veo for AI short film 'A Dragon's Tale' — creatoroff · 2026-09-17
- They thought it was an island… until it opened its eyes — Plastic-Tart-8262 · 2026-09-17