Gagandeep Reehal

The MCAP Dilemma Because Your Robot Data Has Three Jobs

MCAP is not a bad training format. It is a recording format, and most of these arguments are two people describing different layers of the same stack.

Aug 3 • Gagandeep Reehal

5

July 2026

TOPS Is a Vanity Metric

Peak compute tells you almost nothing about how your robot will actually perform

Jul 7 • Gagandeep Reehal

4

Why TensorRT Is Not Your Biggest Latency Problem

From Camera DMA to Control Commands: Where your milliseconds really go?

Jul 5 • Gagandeep Reehal

9

1

The Most Expensive Engineer in Robotics Is Excel

On the unkillable spreadsheet quietly running every robotics company.

Jul 2 • Gagandeep Reehal

4

June 2026

Your Robot Is Fine. Your Data Infrastructure Isn't

The bug that doesn't look like a bug - until you've lost six months, and the unglamorous work which determines whether your robot ships

Jun 30 • Gagandeep Reehal

6

The Book That Asks the Right Question and Then Flinches

Why the Poor Don’t Kill Us - Manu Joseph

Jun 22 • Gagandeep Reehal

5

Default Yes.

The resource founders spend without noticing.

Jun 22 • Gagandeep Reehal

3

1

April 2026

Delusional Optimism - Balancing founder’s superpower and self-sabotage

The same cognitive distortion that builds companies can also quietly destroy them.

Apr 26 • Gagandeep Reehal

3

February 2026

Why the nihilist penguin would make a great founder?

The nihilist penguin isn’t nihilist. It’s realistic.

Feb 8 • Gagandeep Reehal

7

January 2026

The Myth of Product–Market Fit

Why PMF is not found - it’s defended continuously

Jan 27 • Gagandeep Reehal

5

1

Don’t Copy Winners

Why most advice is often survivor bias with good marketing?

Jan 16 • Gagandeep Reehal

7

1

3

AI Won’t Control Narratives. It Will Inherit Them

Why AI-generated content may narrow narratives long before it manipulates beliefs.

Jan 9 • Gagandeep Reehal

6

1