AI is Jazz
Data management evolved like rock and roll — structured, formulaic, predictable. AI evolved like jazz. Most people don't get it. Some people hate it. But to play jazz well, you have to understand the fundamentals. Otherwise it's just noise.
Future-Proof Your Life Science Commercial Data Before Launch
Bad data strategy doesn't only create messy dashboards. It limits what the business can build, automate, govern, and trust later. A conversation on what pre-commercial pharma teams get wrong — and how to pressure-test the foundation while decisions are still flexible.
How to Build an Email Agent That Classifies, Summarizes & Routes — Step by Step
One of the most common early AI agent use cases isn't flashy — it's email. Classify, summarize, route. It sounds simple. The implementation has more moving parts than most expect. Here's exactly how to build it: trigger, classify, extract, summarize, deliver.
Die Hard AI
Hans Gruber's plan was precise. McClane's intervention produced a dozen bodies and millions in damage. Nobody did the math. The AI hero narrative creates a social cost for skepticism. The technology stops being assessed and starts being defended. In Nakatomi Plaza it cost lives. In AI it's costing something harder to count.
Why Your Claude Cowork Sessions Get More Expensive Over Time
Every time you send a message, Claude re-reads the entire conversation from the beginning. A 30-message session costs significantly more at turn 30 than at turn 5. Context rot is real — a bloated session doesn't just cost more, it produces worse results.
Everything I Need to Know About AI Economics I Learned from The Sopranos
You're only as good as your last envelope. When we automate out a layer of the organization, the work doesn't disappear — it migrates up to the most expensive people. The delegation layer we eliminated was load bearing. Christopher didn't need AI. He needed a project manager.
AI: Welcome Back, Humanity
Coding saved me after a PTSD diagnosis — a place where the noise of the world couldn't reach. But medicine has a dosage. AI didn't take anything from coders. It just removed the hiding place. And for a lot of us, that's exactly what we needed.
Every Hand Worth Shaking
At an AI conference in San Francisco, every vendor wanted the pharma meeting. Most had no idea what they were actually being asked to do. The ones who lead with buzzwords tell you they haven't done the work. The pitch is the audition. Most are failing it before they open their mouths.
The Case for Waiting
In pharma, if you're not first you're last. Except in AI, where first might mean you pay for everyone else's mistakes. There are no established best practices for AI in regulated industries. The AI companies knocking on your door need this engagement more than you do.
Vibe Coding Isn't the Risk. Shipping Is.
Vibe coding is like a good first date. It looks great, it works, and it'll tell you it's ready for production. It's not. Flat JSON files, no rate limiting, no auth isolation. Don't walk it down the aisle before you know what you're dealing with.
The Bottleneck Isn't Building Anymore. It's Deciding.
For most of product development history, the hardest part wasn't ideas — it was validation. AI lowers the cost of being wrong. You can now stress-test an idea without assembling a full team. The bottleneck was never building. It was always deciding.
What Happens When Scarcity Is No Longer the Driver of the Economy?
Economics is built around scarcity. AI is introducing artificial abundance in domains previously defined by natural scarcity. When the floor of competence rises across every knowledge domain simultaneously, value shifts toward something that can't be replicated at scale: judgment.
The Solo Economy: Introducing C2C
AI is removing the constraints that defined who could create. Skill gaps narrow. Time becomes less binding. Scale no longer requires capital. The question is no longer what company you work for. It's what you've built that works without you.
Go Home, SQL WHERE Clauses. You're Drunk.
15+ years of SQL and the first instinct never changes: add a WHERE clause. It's fast, it works, and it's technical debt with good syntax. One well-designed lookup table fixes what a hundred WHERE clauses can't — and the next developer will actually understand your work.
The Craftsman & the Machine
Coders are like craftsmen — they insert their personality into their work. AI hasn't changed that. What it changed is how the craft is delivered. The blacksmith who picked up the power tool became more valuable than ever. The one who panicked was replaced.
We started 2516 Technologies because we kept seeing the same problem. Pharma companies. Defense contractors. Financial institutions. All sitting on years of valuable data — and none of it ready for what AI actually requires.
The models aren't the hard part. The foundation is. That's what we do.
Follow on LinkedIn
Voices worth
listening to
We're curating a set of guest contributors — practitioners, operators, and domain experts who think carefully about data, AI, and the decisions that matter in regulated industries.