Raw LLM Responses

I respect your opinion and like your content, but there is a lot of things you said that are wrong. I built a 37 app, enterprise grade monorepo with ChatGPT Plus, Google Gemini 2.5 Pro, Google Vertex Ai, DeepSeek Coder V2 and Claude Opus. It took me around 3 months, and I spent roughly $150 dollars. I didn't touch a single line of code. The average token capacity for each platform is on average 128,000 if you use the premium plans which cost on average $20 a month. The free versions have a much smaller token capacity, because they are free, and generative Ai platforms are resource hogs, so they can't expand capacity for non paying users unless they shell out billions in operational expenses. All the renown generative Ai platforms can fully understand large code bases, depending on how many files you feed it at a time. Google Gemini has a whopping 1,000,000 token capacity in both the Flash and Pro versions of Gemini, so you could feed it a whole app, and it would understand how to refactor it if you prompted it to do so. It can do it in minutes (I can give you a demo you can show your users if you want). You might need to update this video, because in the 4 months since it came out, almost every mainstream generative Ai platform has advanced beyond comprehension.