Home › Hard problems › India

India: hard problems at India's scale

Useful AI on a ₹10,000 phone, offline

Solo: strongOne person can run a real test of this in about 8 weeks.

demo ✓a flagship phone:where demos runAI model1–3B parameters₹10,000offline, in Hindimost Indianphones arethis oneMemorythe model barely fitsBatteryevery answer drains it~~Heatit gets hot, then slows down
On-device AI demos run on flagship phones, but most Indian phones are budget models, where memory, battery and heat are the limit.

Why it is still open

Memory, battery and heat limits. Most "on-device" demos run on flagship phones, while most Indian phones are budget models.

An experiment one person can run in 8 weeks

Run a 1–3B model offline on a budget Android phone for one task (form filling in Hindi). Measure speed, battery used per answer and accuracy.

Verdict: Solo: strong

Easy to demo, and the audience is hundreds of millions of phones.

The solo test behind the verdict

  1. Can it be measured? There is a number that shows you did better, such as an error rate, a catch rate or a cost per answer.
  2. Is it cheap to try? It runs on a laptop, free open models and open data, or under about ₹50,000 of rented GPU time.
  3. Is there a public yardstick? A benchmark or dataset others use, so your result can be compared and believed.
  4. Would a small win get noticed? A lab, a funder or a customer would care about a 10% improvement, not only a full solution.

Four yeses means Solo: strong. Two or three means Solo: possible. Fewer means Follow, don't build.

PreviousReading Indian documents: land records, forms, handwritingNextAddresses that machines can't read

All 79 hard problems · Working on this, or know of a result we missed? basant@cleandataengineering.ai