Epistemic Debt: The Hidden Cost of AI Speed
AI makes drafting cheap. The bill arrives later as review load, churn, security risk, and epistemic debt. The teams that win treat AI like an intern: fast, useful, and never unsupervised.
...
AI makes drafting cheap. The bill arrives later as review load, churn, security risk, and epistemic debt. The teams that win treat AI like an intern: fast, useful, and never unsupervised.
Autocomplete isn’t the real shift in AI-assisted coding. The leverage comes from agentic work, disciplined delegation, and keeping human responsibility where it pays off.
Can you run AI coding assistants locally? Yes - but with caveats. Here's what actually works, what hardware you need, and whether it's worth the hassle.
Behind the affordable AI coding assistants lies an uncomfortable truth: massive investor subsidies, billion-dollar losses, and growing environmental costs. What happens when the subsidy ends?
I was invited to chat with the Packet Pushers folk on 'OpenTelemetry – Getting From Visibility To Observability With Ben Hall'. Looks like I pushed a show that's typically around 40 minutes long, over an hour for my guest appearance. Hope it's a good listen and you learn something new, audience!
At its core, OpenTelemetry (OTel) is a vendor-neutral standard for telemetry. But it also provides a complete end-to-end implementation for generating, emitting, collecting, processing and exporting telemetry data to any support observability back-end.
The observability problem - Why is the system exhibiting this behaviour?
An interesting read about outsourcing and in-house capacity, relevant to us in IT.
A 3-part series published on CodeProject that demonstrates how to construct a complete end-to-end GitOps workflow using Terraform plans, GitHub, GitHub Actions, and Azure.
An alternative using WSL 2, where Docker Desktop is not available to you in Windows