Mike Koenigs uses about ten AI tools at once, treating each as a specialized blade. He defaults to ChatGPT for general work, Claude for copywriting, and GenSpark for multi-step automation. He also uses Grok for uncensored output, Perplexity for research, and warns that Chinese tools like Minimax carry geopolitical risks.
Chapters
Which AI should I use and when? This is a chart that I put together because typically when I am solving a problem, I don't just use one AI. I use about 10 of them simultaneously. And that's one of the ways that I use the Comet Browser very successfully. So my Swiss Army knife is, as you can imagine, ChatGPT. I use it all the time and it continues to get good. Is it the best all the time? No. I have no loyalty to any tool at any given time whatsoever. I also use Claude frequently because it's a really good copywriting and development tool. My second go-to, however, is GenSpark. It is probably the most flexible of all of the tools, and it's considered a super agent, which basically means it automates multi-step tasks. It thinks ahead, and it also creates amazing documents. It writes good code. They just added a design feature. And it's my general purpose go-to for most things other than ChatGPT. Then of course there's Grok from Elon and Team XAI. It is not a censored model, which is often really useful, especially if you absolutely despise arguing with these stupid AIs that decide that something is not okay for whatever reason it may be. That's up to me, not it, especially if I'm paying for it. That's my opinion. Perplexity is a good research tool. Manus is my second favorite of all the super agents, unbelievably powerful. And then there's a few you may not have heard of before, Minimax, Abacus, and Qimi. They are from China. Along with Chinese technology comes Chinese risks. However, these tools are so powerful. Again, I don't have any loyalty to any tool. I just want my stuff done.
