E53: Apertus AI Disaster (2 of 2) - Academic Theater - Why Swiss Companies Still Use ChatGPT
Key Insights
Breaking down the operational disaster:
- The Access Paradox: "Open source" but needs Hugging Face account | Legal indemnification clause | Doesn't work with llama.cpp | Documentation "coming soon"
- Context Length Crisis: Apertus: 65k tokens | Gemini 2.5 Pro: 1M tokens | Llama 4: 10M tokens | Can't even fit one annual report
- The $50M "Reproducibility": 10M GPU-hours on GH200 chips ($3-5/hour) | 15 trillion training tokens | Nation-state budget required
- Safety Nightmare: NO content moderation | Provides bomb instructions | Planned for SCHOOLS | Wild West once downloaded
For access to the full transcript, join our community!
get started



