It's Not the Tech, It's the People and the Relationships that Matter

It is likely that sometime within the next six months OpenAI will make a model available known as Strawberry or ProjectQ that is capable of higher-level reasoning, including being able to do high-level math problems without relying on training data. This could potentially allow it to complete tasks autonomously and generate high quality synthetic data that it can use to train its next model, potentially called “Orion.”

Many argue that OpenAI needs to do this, as there is competitive slippage with Meta, which has attracted 400 million monthly users and Anthropic’s Claude, which appears to have stronger reasoning skills than ChatGPT-4o.

While releasing more advanced reasoning AI at this stage will obviously help OpenAI compete, it is unlikely that such technology is what will really enable it to compete long-term. One thing we’ve learned over the last two years is that the major companies don’t retain monopolies on model capabilities, and efforts by companies such as Meta and Mistrial to open source their developments only make this more likely. Currently, all of the models (ChatGPT, Claude, Gemini, Mistrial (France), Quin (China) etc.) have similar capabilities.

Eventually, it is likely that all the major companies will eventually release AIs that can engage in natural language conversation (as they do now), demonstrate high level reasoning (which appears imminent), engage in hierarchal planning, and demonstrate common sense. Some people doubt that AI will be able to do all of these things, but that really isn’t the critical point. The critical point is that any new technological development is likely to be a momentary edge. Everyone soon gets what someone develops.

So, then what matters? How do companies compete? Your business is unlikely to have, at least for long, a significant technology advantage because it is working with one AI model company vs. another based on the technology.

Companies that use AI will have to compete based on the products they develop with the AI capabilities (call center tech, for example). Companies that manage AI models and support clients will have to do so in a way that integrates company data in a way enables those clients to tailor recommendations, enhance satisfaction and loyalty, develop advanced data analysis for market insights, and inform product development to meet demands and manage supply chains. I’m sure those in business could go on and on here.

In the context of education, think of the “wrappers.” I was initially skeptical of these products, and I still am a bit. Early on, I wondered how someone could possibly be selling school districts wrapped-up AI tools that could generate lesson plans, quizzes, vocabulary lists, slide presentations, and provide feedback on essays when one could simply do that for free with ChatGPT3.5 and other tools such as Claude Artificacts (which can do all of these quite well). It’s also quite simple to build your own bot trained on the data you want.

But the answer is simple: These companies have built relationships with teachers and administrators. They have demonstrated trust and caring. They’ve also built dashboards that make these things easy to use (that anyone with some modest programming experience can also build) so people could use the tech without coming up with their own prompts. So, they’ve developed an easy-to-use product on top of existing technology and built relationships to integrate it into schools. And since these systems largely reinforce existing instructional models, they don’t have to challenge any current educational grammars to sell the products.

Anyhow, I digress a bit, but my large point is that it’s not really about the tech; it’s about the products and the relationships. The “AI companies” aren’t going to win long-term by having a unique AI capability. EdTech companies aren’t going to win long-term by having some unique product capability.

I think those who win long-term are -

• AI companies who can use their AI models to support the needs of specific businesses, including effectively integrating/their business customers’ data and who have not only the technology but the people to make that happen.

• Product companies that develop products that are useful to people and that are enhanced with AI in a unique way.

• Product companies that have strong personal relationships with their clients. Sorry, but it’s not expensive or complicated to duplicate any of the educational wrappers on the market. What will determine success is building relationships with teachers and school administrators, as well as having the mojo to grease the professional development needs for continued use and evolution.

• AI & product companies that can exploit any new innovation for a short-term gain. That will depend on having what one may think of a strong existing AI “pipeline” with an AI model developer (or some really strong in-house capacity).

Note that the first three of these are really about people, not about technology. And even the last one is also about the people to a degree.