Why We Need Competition in AI

The relentless race for cutting-edge AI technology spans all industries – and isn’t slowing down. In the short time between writing this sentence and publishing it, AI will undoubtedly have made remarkable strides, thanks to the fierce competition driving innovation. As the CEO of a data startup, I’m well aware of the intense pressure to have top-tier AI technology. This challenge can feel particularly daunting compared to the significant investments made by industry giants, especially in the intricate domain of Large Language Models (LLMs).

Thankfully, President Biden’s recent AI Executive Order is aimed at fostering investment and competition in this space. By promoting a fair, open, and competitive AI ecosystem, SaaS platform businesses like HData are able to do what we do best: building sophisticated frameworks around LLMs in order to do the work that our customers need done.

We are never going to build our own LLMs. Instead, we embed multiple LLMs in a framework that adds pre-processing, supplemental analysis, access to one-of-a-kind source data, and specific user experience.

For example, HData's frameworks allow the embedded LLMs to answer questions about energy regulation without hallucinations, and to supply direct citations to the underlying subject matter, so our customers can understand where the answers came from. LLMs by themselves cannot do those things, but with the addition of the framework of technologies we build around them at HData, our customers now can.

If our customers tried to use "raw" LLMs without the assistance of frameworks built around them, then they would risk incorrect answers to their queries and be unable to see how the LLMs came up with answers. Our frameworks solve those two problems, which means our Regulatory AI technology is perfect for the regulatory work our clients engage in.

SaaS platforms like HData’s are fit-for-purpose: the framework allows us to serve a specific user need. In our case, that’s the need to manage and understand energy regulatory information.

Our customers need to manage and understand regulatory filings and regulatory agencies' announcements, accurately, in less time. In order to address the individual needs of our clients, we want to have a meaningful choice between multiple large language models to incorporate into our technology stack.

It’s this necessity of choice that drives our work at HData. Each of our clients, while they may fall within the same type of work or industry, have unique questions that require answers to be not just accurate, but time-sensitive and citable.

Competition within the AI industry is essential to harness the true power of AI across the economy.? We welcome the Biden Administration's push to prevent market dominance so that innovation can flourish. It is not out of the realm of possibilities for the LLMs market to commodify, just as the market for cloud computing did in the 2000s.