Skip to content

On-Device AI

Some of the most interesting AI now runs without a round trip. This topic follows local inference and its limits.

Local inference on phones, laptops, and dedicated hardware, and the privacy and latency case for it.

5 stories

A Copilot+ laptop showing Windows Studio Effects beside the NPU graph in Task Manager.

Consumer AI Hardware

A Copilot+ NPU Only Matters When Your Apps Use It

The Copilot+ badge says a laptop has a capable neural processor. It does not guarantee that the AI features you use will run locally, respond faster or extend battery life.

Devin Oyelaran · August 9, 2026 · 8 min

An RTX-equipped Windows laptop transcribing a 60-minute meeting in Buzz beside a digital audio recorder.

Consumer AI Hardware

When Your Laptop Can Replace Paid Transcription

Run one 60-minute recording locally before canceling a transcription subscription. Processing speed matters, but speaker labels, language support, battery use, and privacy usually decide the result.

Devin Oyelaran · 8 min read

Support ticket text beside structured fields for issue type, order number, summary and selected internal tool.

AI Industry & Models

Use a Smaller AI Model Until the Accuracy Breaks

Smaller models can classify, extract and route routine work with less delay and expense. The useful boundary appears only when each task is tested separately.

Tobias Lund · 7 min read

Other impacts

One story a day

The story of the day, in your inbox

One real story about AI each morning — no hype, no alarm, just company for the road.