For the last few moths I have been exploring building a local LLM specifically trained on data from pupils with additional SEMH needs and also SEMH frameworks and approaches. Yesterday I read an article from Ray Dalio discussing building an AI Clone of himself that is trained on his 40 years experience in finance. I think what I and Mr Dalio are building are the same, but with a VERY different budget.

From School SENCO to Global Finance Guru

Ray Dalio, the mind behind Bridgewater (one of the most successful hedge funds ever), is a titan of finance. He’s now launched his next big venture: an ‘AI Clone’ of himself.

What is he doing? He’s taking the entirety of his proprietary knowledge—over 40 years of market analysis, investment principles, and management memos—and training an LLM on it. He’s essentially packaging his unique expertise and selling access to it, promising to ‘coach’ people and scale up his personal influence infinitely. He writes about it himself here.

The takeaway here is massive:

1. The Tech is Proven: The fact that a world-class business leader is betting his reputation and next enterprise on this exact model proves the approach is technologically sound and capable of generating significant, measurable value. Training an LLM on a specialised, proprietary dataset to create an expert tool that works.

2. My SEMH Project is on the Right Track: If this model can scale four decades of financial expertise, imagine the consistency and efficiency it can bring to assessing complex SEMH needs and streamlining SEND support for teachers! It confirms that creating an intelligent, focused assistant based on expert data is the future.

This validation is incredibly encouraging and gives me confidence that I am building something truly useful and impactful for the education sector.

---

The Stark Divide: Public Value vs. Private Profit

However, Dalio’s success immediately highlights a colossal problem for the public sector, and specifically for those of us working in education: the monetisation gap.

Dalio can sell his AI Clone because his customers (investors, business people) can easily calculate a return on investment (ROI). They can afford to pay for an unproven innovation because the potential financial gain justifies the risk. The tool becomes a profit centre.

Now, let’s look back at our schools:

• We Create Societal Value, Not Financial Profit: The true value of a tool that supports SEMH pupils is better outcomes, less stress on teachers, and reduced long-term public service costs. This value is immense, but it doesn’t fit into a tidy commercial model.
• The ‘Customer’ Cannot Justify the Spend: How can a school with a ‘tight’ budget justify paying a high commercial price for an innovation, however beneficial, when they cannot see a direct, immediate financial return? They are often forced to wait until a tool is ‘proven’ and its cost significantly reduced.

This means that incredibly necessary tools—like a local LLM that could revolutionise SEMH support—are often not built by the big commercial players because they cannot make enough money.

The irony is not lost on me: the very principle Dalio is using to generate revenue is the same principle that could save our public services money and, more importantly, transform the lives of countless vulnerable young people.

This dichotomy reinforces the need for bespoke, often locally-driven, public sector innovation. We can’t wait for Big Tech to solve our problems if there isn’t a massive profit to be made. We need to be building these tools ourselves, supported by strategic public funding, because the value created—better outcomes for our pupils—is priceless.

I am aware that it is significantly easier to identify a problem than fix it. I do not have the solution, but I hope by sharing this, it reaches someone that maybe does…