https://www.luiscos.io/blog/ Writing about AI safety, alignment, mechanistic interpretability, and security. en-us Tue, 09 Jun 2026 00:00:00 GMT https://www.luiscos.io/blog/caltrops/ https://www.luiscos.io/blog/caltrops/ Sat, 23 May 2026 00:00:00 GMT Plant a useful reasoning habit in a teacher model. A distilled student inherits the habit form without the substance, fails confidently, and leaves a detectable fingerprint that survives paraphrase. https://www.luiscos.io/blog/uinference/ https://www.luiscos.io/blog/uinference/ Sun, 25 Jan 2026 00:00:00 GMT Can frontier AI inference run on a radically minimized stack while protecting model weights against nation-state adversaries? An seL4-backed Weight Enclave prototype. https://www.luiscos.io/blog/openwild/ https://www.luiscos.io/blog/openwild/ Sun, 25 Jan 2026 00:00:00 GMT An open conversational dataset capturing real-world interactions with current-generation language models. Inspired by WildChat. https://www.luiscos.io/blog/mechanistic-watchdog/ https://www.luiscos.io/blog/mechanistic-watchdog/ Wed, 15 Jan 2025 00:00:00 GMT A safety layer that monitors a language model's internal activations and halts generation before harmful content is produced.