Because Sentient has always emphasized open source, they are often asked about "open source versus closed source" during interviews and research. These two are not about which one is better or worse, nor are they about the distinction between liberal arts and science students.

In the short term, large companies like OpenAI and Google set product rules; in the medium term, the open source community defines ecosystem rules; in the long term, the true rule-makers may be at the interface layer and protocol layer.
Currently, the advantages of closed-source companies are very tangible: strong model capabilities, mature product experience, fast delivery speed, and hassle-free enterprise adoption. Therefore, in the near future, closed-source companies are likely to hold the most influential voice.
But on the other hand, open source is not just about ideals; it is a balancing force in the real world.
Stanford AI Index 2025 explicitly states that the performance gap between open-weight models and closed-weight models is rapidly narrowing, shrinking from 8% to 1.7% on some metrics. This means that in the future, only a few companies will be able to develop Agents, and more and more teams will be able to create usable systems within controllable costs.
So I don’t believe it will ultimately be "OpenAI dominates everything," nor do I believe in a "complete victory of open source idealism."
More likely, the capabilities of foundational models will continue to concentrate, but the upper-layer agent ecosystem, tool invocation, identity, payments, collaboration, memory, and auditing interfaces will become increasingly open.
By then, who sets the rules may not be the company with the most parameters, but whoever controls the standards for agent interoperability, the workflow entry points, and the developer ecosystem.
The true battleground of the future is not just the models, but the operating systems built on top of them. That’s why strategies like Sentient’s focus on open agent economy and open infrastructure make strategic sense.