Our Argument for Open Source AI.

We all know we’re very early in this AI game. When these societal transformations come about in a capitalist market, it’s almost always a shoot-out among well-financed companies to find the edge. This means that the new tech — like AI — is market driven, meaning that these companies will rapidly evolve their products to drive any market value they can, in as fast a timeframe as possible.The only way for them to do this is to control their product.

Today, that means that we have myriad competing LLMs that integrate with their existing core products: Google and Gemini integrating with Google products; Open AI embedded into Microsoft 365; Anthropic’s Claude is closed-source; and so is the advanced Mistral LLM.

Of the largest language models, only Meta’s LLAMA is truly open source. I have to assume that Elon Musk’s xAI will be open source since he’s been a thorn in Sam Altman’s side for not making Open AI an open source platform like its original mission.

Open source AI models that are as powerful as paid versions are paramount to a more equitable future. These closed models — as amazing as they can be — are a problem. The current trajectory of AI under the current closed models means that people with the means can acquire the best AI that money can buy.

AI is unlike computers in the sense that AI provides the user with potentially unlimited intelligence and capabilities. The question is: who deserves unlimited intelligence? Who “gets” to have their job automated and yield the fruits of a machine’s labor? Only the rich? Only those with means?

Imagine Google search not being open to everyone. Imagine that only people with means can search the Internet. We wouldn’t do it, would we? The market decided at some point in the 90s that Internet search was too important for it not to be available to everyone. Of course, that became a giant addressable market that Google monetized through AdWords. But the capabilities of search was free to the user.

Closed AI architectures are only good for the companies that create them. And these gains may be short-lived. Interoperability becomes unmanageable and inflexible. AI should be considered a basic human necessity if it truly does create a new world of productivity, abundance, and human advancement.

If only people, companies, countries, and communities with the means can utilize AI, then we are headed into a dark future where the haves propel forward and the have-nots are left behind. This is not a future rooted in ethics, goodness, and possibility.