EngineeringMay 12, 2026
Why we run Whisper on the laptop, not in the cloud
The architectural and commercial argument for keeping inference local — and why it changes how IT teams evaluate AI tools.
Case studyApril 28, 2026
Fee earners saved 6.4 hours a week. Here's the math.
A breakdown of a 90-day pilot at a 200-lawyer firm: time saved, errors avoided, and what changed in practice.
SecurityApril 4, 2026
Passing an enterprise IT review in under 48 hours
The questions security teams actually ask, and the documents that answer them upfront.
EngineeringMarch 18, 2026
Whisper v3 on Apple silicon: latency benchmarks
End-to-end cold start, warm start, and first-token timings across M1, M2, and M3 chips.