All posts Technology

The roadmap is less hardware, not more

MKMalli KosuriMar 20266 min read
In short

Component prices climb and an accessible price is core to the mission, so the pen's future is subtraction: every part must repeatedly justify itself against the question of whether software plus the remaining sensors could do its job.

Accuracy is the fixed constraint, the digital twin is the proving ground, and nothing is removed until simulation and hardware trials show the reconstruction does not care.

The short version. The full post has the detail and the why.

In this post

  1. Why subtraction is the strategy
  2. Software eats sensors, carefully
  3. Accuracy is the constraint, not a trade
  4. Rehearsing subtraction in simulation

Why subtraction is the strategy

For a product meant for every school desk in India, the bill of materials is destiny: every component is multiplied by every pen forever. Electronics prices have been unkind, and the honest responses are either raising the price, breaking the mission, or engineering the cost out. So the roadmap treats part-count as a metric to shrink, generation over generation, the way others treat features as a list to grow.

Software eats sensors, carefully

The pattern that makes subtraction possible: information one component provides can often be recovered from the others plus better mathematics. Orientation leans a little less on one sensor as fusion of the rest improves; a dedicated function migrates into an algorithm. Each such move deletes a part, its board space, its assembly step and its failure modes. The discipline is moving only when the recovered signal is provably as good, which is a mathematics problem with a clear pass mark.

today software the roadmap
Generation over generation: a denser board becomes a sparser one, with algorithms absorbing the work of removed parts, and accuracy gating every removal.

Accuracy is the constraint, not a trade

One rule shields the whole exercise: accuracy is not on the negotiating table. A cheaper pen that reads a child's writing worse is not a cheaper pen, it is a broken promise with a lower price tag. Every proposed removal is scored first and only on reconstruction and recognition quality. Plenty of tempting subtractions have failed that scoring and stayed in the pen. That is the system working.

Rehearsing subtraction in simulation

The digital twin, our software model of the pen's sensors and physics, is where removals audition. Disable a component in simulation, replay thousands of writing sessions, measure what degrades. Survivors graduate to real prototype trials with the part depopulated. Only after both gates pass does a subtraction reach production. The goal a few generations out is a pen with fewer parts than today that reads handwriting better than today, and both halves of that sentence are the point.

Key takeaways

  • Part count is a metric to shrink; the bill of materials is multiplied by every pen forever.
  • Software plus better fusion can absorb components, but only with proof.
  • Accuracy is the immovable constraint; failed subtractions stay in the pen.
  • The digital twin rehearses every removal across thousands of sessions before hardware does.

Anyone can add a part to fix a problem. The craft this product demands is removing one without creating any.

Read next