Why AI team coordination needs more friction
Automating every quick question makes tasks faster and teams weaker. What the research on informal communication and AI burnout shows, plus three ways to put useful friction back into the workflow.
Category: AI
Why my personal AI assistant ditched ICS for the Calendar API
Part two of the Fernao build log: the ICS calendar feed needed five minutes to generate a schedule, so I moved to the Google Calendar API and got it down to twenty seconds. The new Task Breaker module splits oversized tasks into 20-minute subtasks.
Cutting a 220,000-token document down to 4,000 with K-means
Part 2 on summarizing a 200,000-word handbook: validating K-means clusters with UMAP and silhouette scores, picking the chunk nearest each centroid, and where the final reduce step still drops topics.
Why byte-level models beat LLMs for cross-script name search
Levenshtein and Soundex break the moment a query crosses script boundaries. Notes from a sanctions screening build on byte-level transformer embeddings, ANCE hard negative mining, and where CJK still causes trouble.
Zero-shot classification instead of regex for messy data
Regex and K-means both fall apart on short free-text data where the same idea gets phrased three different ways. Zero-shot classification with a local model like Gemma 2 handles the paraphrasing, and this is the Ollama pipeline I ran over several thousand security annotations.
Scaling RL with Approximate Solution Methods
Tabular RL breaks the moment the state space grows past a toy grid. Approximate Solution Methods swap the lookup table for a weight vector and semi-gradient updates, so one state teaches you about every state like it. Includes where linear models still beat a neural net.
How a local SLM fixed my nightly CI/CD pipeline
Twenty-three pipeline failures in six weeks, none of them caused by my code. This is how I replaced GPT-4 with a seeded Qwen2.5 on Ollama, plus the GitHub Actions caching and Pydantic validation details that made it stick.
Agentic AI supply chain: finding the delays nobody owns
A logistics director with all-green KPIs and 18% late shipments does not have a data problem. Scoped AI agents watching the TMS can tell you which handoff actually broke, and tell you on the morning it happens.
Lasso regression and the geometry that stops overfitting
A house-pricing model with zero training error valued a suburban shack at $2.4 million. This is the geometry behind lasso: centering the data, the L1 diamond, and why landing on a corner zeroes a coefficient.
Causal inference analysis: what actually caused the spike
Naive before-and-after averages said the London tube strikes lifted bike usage by 5.5%. A two-way fixed effects model with clustered errors put it at 3.95%. This is why the second number is the one to trust.