Founder of Friday AI • 2m
🚨 Open AI is an Wrapper👀🤯 Hot take, but let’s break it down logically: OpenAI is not a full-stack AI company — it’s a high-level wrapper over Azure and NVIDIA. Here’s why that matters 👇 🔹 1. Infra Backbone = Microsoft Azure Almost 90%+ of OpenAI's training & inference happens on Azure superclusters. Without Microsoft’s cloud muscle, the scale of ChatGPT or GPT-4 would be impossible. ➡️ OpenAI rents brains from Microsoft and calls it a mind of its own. 🔹 2. Compute Core = NVIDIA H100s Those "intelligent" answers? Powered by stacks of NVIDIA H100s and A100s. No silicon = No LLM. ➡️ OpenAI doesn’t own the factory. It rents the engines. 🔹 3. Business Model = API Layer What’s OpenAI really selling? 👉 APIs on top of rented compute. It’s brilliant middleware: take raw GPU power → train/fine-tune models → wrap in APIs → serve to devs, startups, and enterprises. ➡️ Think Stripe for Intelligence. 🧠 The Real Game? > Vertical integration = owning chips (like NVIDIA), infra (like Azure), and intelligence (like OpenAI). But OpenAI owns only the top — the cognitive interface. In short: OpenAI ≠ end-to-end AI It’s the middleware king of Gen AI. Still legendary, still powerful. But let’s call the architecture what it is: A wrapper. A brilliant one. #OpenAI #Azure #Nvidia #LLM #AIInfra #StartupInsights #FridayAI #BuildInPublic #VCLogic
Ask yourself the que... • 4m
Nvidia Isn’t Just Winning—It’s Redefining the Game What if I told you the most important company in the world right now might not be Apple, Amazon, or Google—but Nvidia? Once a niche GPU maker for gamers, Nvidia has now become the oxygen for AI. Fr
See MoreHey I am on Medial • 1y
The AI chip war is coming, with players like NVIDIA, Samsung Semiconductor, Rebellion, OpenAI, Huawei, and Qualcomm. Who do you think will dominate? Many Investment Banking Institutions are investing Billions of Dollars in AI Chips Companies like Ti
See MorePython Developer 💻 ... • 5m
3B LLM outperforms 405B LLM 🤯 Similarly, a 7B LLM outperforms OpenAI o1 & DeepSeek-R1 🤯 🤯 LLM: llama 3 Datasets: MATH-500 & AIME-2024 This has done on research with compute optimal Test-Time Scaling (TTS). Recently, OpenAI o1 shows that Test-
See MoreDownload the medial app to read full posts, comements and news.