Notes on data, AI, IT
and security
No marketing fog. The way I think about real problems with founders and managers.
Real-time data and right-time data: the difference and why it matters
Not every task requires real-time data. Getting this choice wrong costs money and complicates architecture without benefit.
Vendor lock-in: measure the cost of leaving, not the cost of entry
The question when adopting a platform is not only what it costs to get in. It is what it would cost to get out - and whether you can honestly answer that before signing.
Feature engineering is a business decision in disguise
The variables you feed into a machine learning model are not a purely technical choice. They encode assumptions about your business that deserve explicit review.
Collaborative robots: the economics of a cobot for managers
Cobots are not new, but in 2016 their cost and ease of deployment have reached the point where the economics conversation has become practical.
Who owns the data pipeline when the answer is nobody
In most companies data pipelines are built by whoever needed the data, owned by nobody, and relied upon by everyone. That is a systemic fragility, not a technical problem.
How architecture changes after large breaches and trust failures
Yahoo, LinkedIn, Dropbox - 2016 showed that breaches happen to everyone. Here is what this changes in how companies should think about trust architecture.
On-call rotation is a management problem, not an IT problem
Unstructured on-call duties burn out engineers and leave incidents without clear ownership. The fix is not a tool - it is a set of decisions that only management can make.
The enterprise case for a password manager, made simply
After another year of major credential breaches, the argument for a corporate password manager is no longer mainly technical. It is an operational risk argument.
Chatbots: between the hype and the first practical use
In 2016 everyone is talking about chatbots. Here is where they actually work and where they are a marketing promise.
What Pokemon Go's outages teach about location data at scale
Pokemon Go is not a business application, but its infrastructure story in the summer of 2016 is a real lesson in what location data at scale actually costs.
Deployment frequency as a safety metric, not a speed metric
DevOps changes not only how fast code ships, but how risky each change is.
PostgreSQL JSONB: when you do not need a separate NoSQL database
Before adding MongoDB or another document store to your stack, it is worth checking what PostgreSQL's JSONB type can already do - and where it genuinely runs out.