Casasol

The idea fit on an index card. A small model on an iPhone answers questions about buying property in Spain — in English, Polish or Spanish — the way a colegiado would: gives the Spanish term, cites the tema and the article, and refuses to make the decision for you. The facts never live in the model. They come from retrieval over a study guide one of us wrote this summer while qualifying as a real-estate agent, about 300 KB of it, on the device. Only the manner goes into the weights: which language to answer in, plain text, cite, hand off. Facts in retrieval, form in the weights. Nothing leaves the phone. ...

In June, after the US government suspended two Claude models for every non-US user overnight, we wrote about what CasaSol experienced: nothing. No API call to interrupt, no data in transit to retain, no dependency to lose. We closed that post with a promise — Chronos experiment 018 would stop asserting that and start testing it. Three scenarios: stop the inference daemon, delete the model weights, cut the network entirely. The hypothesis was that all three degrade gracefully with zero data loss and configuration-only recovery. ...

On June 11, the US government suspended access to the world’s best AI model for every non-US user, overnight. Here is what CasaSol experienced.

Human project reviews are slow. The bottleneck is not judgment — it is context reconstruction. Before you can criticise anything, you spend twenty minutes remembering where you left off. The question we asked: can a local 26B model serve as a recurring adversarial QA critic that catches real problems, not just surfaces obvious gaps? Enter Experiment 009. The Setup Two critics. Same project context. Fixed evaluation schema. No collaboration between runs. ...

If every company can be a Palantir now, how do you test that claim? Generating ideas is not difficult. The best frontier models make strategic brainstorming surprisingly cheap. But a well-formed idea is a long way from execution. The real world is messy, chaotic, and constantly adapting. So the only honest test is to build the thing. The solution is to bootstrap a local Palantir and watch what happens. Local, in this case, means the Costa del Sol — known for its climate, its golf courses, and its expensive real estate. The region has between 2,000 and 2,500 active real estate companies. Marbella is the undisputed centre of gravity. The market spans the full spectrum: global brands with multi-office setups, local boutiques that have operated for twenty years or more, and independent agents — mostly property finders — collaborating with larger agencies through shared network databases. ...

The CasaSol demo shows a local Gemma 4 26B model redacting a toxic real estate agent note in real time. The implicit claim behind that demo is that the output is GDPR-clean. An anecdote is not an experiment. One demo run is a marketing moment. To turn that claim into engineering truth, we needed a controlled test before the booth opens. Enter Experiment 006: The Redactor Fidelity Test. The Redaction Contract Redaction is a contract. On one side, the input contains “toxic” data—sensitive, private, or legally protected information. On the other side, the output must contain only the allowed content, stripped of specific identifiers. ...