William Gibson once observed:
The future is already here—it is just unevenly distributed.
We hold a complementary truth:
The future of engineering is already here—scattered across models, tools, frameworks, and APIs—awaiting the architect who will weave it into coherence.
Notes on our mission
As we started our journey to build an AI mechanical engineer, we stumbled upon a realization: our physical AI standard text-based coding workflows only solve part of the problem.
Every mechanical engineer—or any engineer—may appear capable of accomplishing many things, but their abilities are built upon thousands of human beings who have collectively advanced human knowledge through small contributions and improved our collective understanding.
We believe that the best way to develop an AI mechanical engineer is to follow the same path that nature during its 4 billion years of quest for intelligence has followed.
Evolution Of Human Intelligence
Human intelligence has evolved over an astonishing span of four billion years, and today, with the advent of machine learning and neural networks, we are essentially attempting to reverse engineer that intelligence.
When we write, we crystallize our thoughts. Training neural networks to reproduce written text means we are training them to mimic the patterns inside our brains that result in those texts/thoughts. This is how reverse engineering works. Through scale, structure, and intent, we seek to unlock the emergent behaviors that approach the underlying mechanics of human intelligence and cognition.
We can already see very established coding abilities emerging through large language models. Since we are reverse-engineering intelligence, any specialized roles we develop in machines that mimic human intelligence will likely replicate patterns of advanced cognitive abilities.
Instead of building a digital engineer that can perform multiple functions from scratch, it's better to build specialized engineers that can perform one task very well. Later, one can integrate them to perform a multitude of specialized tasks, giving the impression of a general engineer.
This is why we split our development work into different specialized roles, which we believe can now be accomplished with the help of human-in-the-loop systems and large language models.
Intersected Intelligence or Intersected Engineering
Intersected Engineering is when different forms of agents, different agents specialized in their own tasks, come together along with human beings and mimic the decision hierarchy and decision-making hierarchies of human engineering teams to accomplish tasks that mimic a human engineer. The foundational aspects or the foundational requirements for that engineering are:
- Visual intelligence
- Logical intelligence
- Specialisation/Focus: Deep knowledge and expertise of a subject.
- Agency: Pursuit until the goal is achieved
- Collaboration among agents and humans
At Ontium, we are building an intersected intelligence that is focused on solving specialized problems faced by engineers.
the current problems that we are solving are manufacturing intelligence, the intersection of mechanical and electrical intelligence