Workers spend a sixth of their time packing — not picking, not sorting. Hands in clamshells, one berry at a time.
Flavor is what consumers care about most in blackberries. Quality is won or lost at packing — bruised berries, bad fills, inconsistent grades.
Farmers lose up to 6% of berries that would have passed quality checks. Rejected by hand-sorting errors, not by actual defects.
Current: $2.50–$6.50 per clamshell. Latest: $3.10. Target: $0.50 — human cost.
120+ fps industrial cameras capture multiple sides of every berry. Full coverage, no blind spots, no sampling.
Real-time grading in milliseconds — colour, size, bruising, mold. Every berry scored before it moves an inch.
A high-speed bat mechanism rejects defective berries off the line instantly. No robotic arms, no grippers — simple, fast, mechanical.
FDA / USDA compliant. Food-grade production process. Built with sanitation in mind.
(40 min)
Sees one side of berry
Packing blackberries led to classical robotics — computer vision, custom servos. The next task will be better selected for embodied AI, though some classical robotics will almost certainly be unavoidable. Imitation learning has matured dramatically since we started, and we're better suited to apply it.
We're not narrowing early this time. We'll explore multiple candidate tasks, iterate fast across each, and commit only when the data says to. Same method, sharper tools, better judgment about where to point them — all gained from solving packing.
In Peru, 1% of the national workforce picks blueberries. But zoom into the towns where they're grown and that number is 30%. This technology will eliminate these jobs. So how do we do it in a way that empowers those workers to find more meaningful work?
In the US, severance packages are standard when jobs are displaced. But these workers are outside that system — in countries where no one offers it. If we displace these jobs, nobody else is stepping in.
3% of revenue goes back to displaced workers as a severance — something quite common in the tech industry. For a company, 3% is insignificant. What matters much more is our growth — as BluPe scales, the fund scales with it.
Upskilling programs sound good on paper. Direct severance is simpler and lets people spend on what they actually need — not what a corporate program decides for them.
We don't have all the answers. If this works, we'll learn what responsible deployment actually looks like. If it doesn't, we'll understand better why.
We're incorporating as a Public Benefit Corporation with a mission of giving back 3% of earnings to replaced workers.
The first investors and early team members will define how this actually works. Who counts as affected? How is money distributed? What role do intermediaries play? Is money enough?
Before the company grows too large, this will get formally written into the company's structure to keep it as a non-negotiable part of replacing labor.