Hey there,

There are a-million-and-one AI tools to choose from. A new one comes out basically every day, if not every hour.

One of the most challenging things for IWAI readers and students is to figure out how to choose – they get fear-of-missing-out, multiplied by analysis-paralysis, with a general overwhelm sprinkled on top.

We've been working on this problem for our students for more than three years now, and the counterintuitive option is:

Don't choose. Instead, lean in to the idea that resilient and antifragile AI enthusiasts develop a platform-agnostic approach to AI.

I'll talk in detail about that today, including how we apply it in all of our programs. (One of our flagship programs, The AI Executive System, is opening later this month – join the early alert list here.)

A case in point about why this is critical: yesterday, Meta released a brand new line of large language models, with the first model being named Muse Spark. I played with it a bit and it seems like it will clearly be a competitor.

You are going to be seeing more and more of this as the big players catch up to OpenAI. We saw Google release the excellent Gemini 3.1 in February. I am already using Gemini to replace the smaller GPT models in my AI agents; it is really good. I expect Amazon and potentially other big companies to hit the market with excellent models in the next year, too.

So, not only do we have 5+ major players (OpenAI, Anthropic, Google, Meta, Amazon) and many small-but-important alternatives (xAI, Mistral, DeepSeek, Qwen), but that just covers language models, not even touching things like video models and orchestration tools.

The speed at which new technology is being released makes the game of "which one is best?" sort of a fool's errand. It is a stark contrast to the old days of software (three years ago