Writing
Essays are longer arguments. Notes are shorter thoughts, many first shared on LinkedIn. Each text says which language I wrote it in; the others are translations made with Saci.
Meaning & AI
- The mystery of knowledge
We've invested like never before in computing to extract knowledge from raw data. These technologies seem to demand that we think better, not less.
- Oriented to data, not only by data
Databases are no longer just backing services. With AI agents as connective tissue, modeling moves up to the layer of meaning.
Data modeling & architecture
- A machine that doesn't know how to remember
The enterprise produces knowledge all day and forgets what it wanted to know. A data model is the minimum shared truth that lets it remember.
- Modeling is not architecture
The medallion architecture organizes a pipeline into layers. It says nothing about how to model the information.
Data as organization
- There is no generic data
Becoming data-driven is not trading intuition for KPIs. Data always has a shape, a source and a life cycle, and it takes a culture to keep it meaningful.
- A default is also a decision
A small card fraud, read with a risk analyst's eyes. Whoever doesn't take on a risk shouldn't make it the customer's default.
- The ROI of a data initiative
Measuring how many fruits you pick is one thing. Measuring whether picking them drains the energy you get from eating them is another.
- Data engineering is building traps
It's not about tools, and not just about data. It's about producing abstractions of meaning, and every abstraction has a cost of understanding.
- All data problems are people problems
Speed and scale can't improve understanding on their own.
- Abstraction never ends
Deriving meaning from information is, for any organization, a process of self-discovery. An endless game, not a tool purchase.
Fundamentals over hype
- Fundamentals outlast waves
Twenty years of innovation and hype in analytics. Fundamental ideas last longer and travel further than any tool.
- There is no magic
The stock and flow of data is a physical thing. Convenient APIs have invisible costs.
Careers in data
- Formal languages and the value that doesn't make noise
Whoever broadcasts gets heard even by those not paying attention. In a world of software and data, formal languages weigh more and more, and nobody becomes a musician just by talking.