AI

I could have started Seasoned Pan’s meal estimator by connecting a camera to a vision model and asking for calories. That would have produced an impressive demo and a dishonest product. Instead, I put four awkward, ordinary food photographs in front of the proposed response format before building the feature: a homemade taco, a frosted … Read more

An automated event card needs a background. The calendar does not have original photography for every venue, and repeatedly generating fake concert photographs would blur an important line in the project. SoCalNomad is partly a photography platform. Synthetic imagery should fill a production gap without pretending to document a place I have not photographed. I … Read more

Three independent sources sounds like a responsible publication threshold. For a major tour announcement, it is easy to satisfy. National outlets, trade publications, label publicity, and entertainment sites all repeat the same announcement. For a legitimate club show or scene festival, three outlets may never cover it. A rigid source minimum can therefore measure publicity … Read more

I noticed the automated newsroom publishing mainstream arena acts while repeatedly overlooking punk, metal, hardcore, indie, and festival stories. The feeds were not the problem. Articles about those scenes were being ingested and passing the Southern California relevance filter. They were dying later because the system had accidentally defined local relevance as popularity inside its … Read more

Food logging fails at intake. The calculations can be perfect and the charts can be polished, but none of that matters if recording a familiar jar, package, or snack requires transcribing a nutrition label field by field. The obvious fashionable answer was image recognition. Point a camera at the food, ask an AI model what … Read more

The automation kept running. Scheduled jobs started on time. Processes exited. Logs continued to accumulate. Nothing obvious announced that the newsroom had stopped publishing. The external AI model it depended on was no longer available. Process Health Was Not Product Health NewsDesk uses language models for classification, entity extraction, clustering, and synthesis. At the time, … Read more

Automated publishers need a way to avoid repeating themselves. SoCalNomad’s NewsDesk already checked whether a recently published story covered the same artist. It also had a title-similarity backstop intended to catch near-duplicate headlines that slipped through the first check. I found ten legitimate stories on the wrong side of that backstop. Entertainment Headlines Share A … Read more

Most image pipelines are designed around delivery. Take a large upload, rotate it correctly, resize it, compress it, and keep the version the application can display efficiently. That is usually a sensible way to build a website. Markit had a different requirement. The photographs were not decoration. They were evidence used to identify an object, … Read more

I started Markit with a tempting automation idea: take item photos and turn them into marketplace listings. That sounds straightforward until the object is old, unlabeled, damaged, unusual, collectible, or valuable. Then the risk moves from formatting a listing to making claims that may not be true. The boundary became clear very early. I want … Read more

The glamorous part of my Markit idea is the AI appraisal. Take photos of an object, identify what it might be, gather evidence, estimate value, and generate listing drafts for marketplaces. That is the part people notice. The less glamorous part is more important: getting the photos into the system without making the workflow miserable. … Read more

The longer SoCalNomad ran, the less the interesting work looked like feature development. It looked like operations. That is not a complaint. It is what happens when a project becomes real. Once people and crawlers can reach it, every change has a blast radius. Every cache has a memory. Every workflow has a schedule. Every … Read more

AI tools are easier to use than they are to manage. The first few wins can make the process feel magical. A model drafts code, explains an error, rewrites documentation, or finds a missing assumption. Then the same model starts making architectural decisions it was not asked to make, changing scope midstream, or treating a … Read more

Most project work disappears because it was never captured in a useful shape. The decision was made in a chat window. The fix lived in a terminal scrollback. The reason for a tradeoff was obvious at the time, then impossible to reconstruct six weeks later. By the time a project is interesting enough to write … Read more

An n8n AI Agent repeatedly called one tool until its iteration limit, leading to a version workaround and stronger repetition guards.

Documentation becomes a liability when current instructions, decisions, and historical evidence compete for authority.

Large workflow payloads, database results, and repeated documentation turned an AI coding session into a bounded-resource failure.

A stale identifier in conversational memory showed why consequential operations need structured intent binding rather than model recall.

A partial-success incident exposed fifty-five WordPress posts whose internal workflow state never reached published.

Why SoCalNomad divided research, editorial judgment, publishing, and marketing into specialized workflows instead of one large AI agent.

An early SoCalNomad experiment separated probabilistic publication timing from deterministic editorial rules, quotas, and safety controls.

AI was never the first hard problem in the SoCalNomad newsfeed. The first hard problem was getting ordinary RSS feeds into a database reliably. That sounds less interesting than clustering, summarization, or automated publishing. It was also more important. If the intake layer is unreliable, every clever downstream step inherits bad assumptions. The 54-Feed Problem … Read more

An early SoCalNomad clustering experiment exposed the need for multi-story article membership and a better separation between deterministic code and semantic judgment.

The first serious SoCalNomad automation problem was not writing articles. It was deciding what did not deserve to become one. The early ambition was obvious enough: collect Southern California entertainment stories from many sources, filter them, synthesize useful coverage, and publish through WordPress. The trap was also obvious once I looked at the data. Entertainment … Read more