Capital de Tokens
Ideas, analysis, and independent articles from the channel and its episodes.
Blog RSSAn Every team completed Anthropic's official training and found something more useful than a bag of tricks: a shared vocabulary for Skills, MCP, the API, and Claude Code. They also exposed the limits of certifying an industry that changes faster than its courses.
What it means for Bubblewrap to be a low-level unprivileged sandboxing tool, how it uses Linux namespaces and mounts, the role it plays inside Flatpak, and why it should not be confused with a complete security policy.
Cloudflare cut the memory footprint of the DNS cache behind 1.1.1.1 by more than half with five Rust representation changes. The result: roughly 100 TB of RAM freed, higher throughput, and lower latency.
Falling Computer Science enrollment and rising demand for quantitative talent are fueling a new thesis: in the AI era, mathematics can be a major advantage. But the strongest profile is not Math versus CS; it is Math + Programming + AI.
Cursor explains how its Router combines a complexity predictor, a task taxonomy, and a budget-constrained optimizer to decide when to use a cheap model and when to escalate to a frontier model.
`--dangerously-bypass-approvals-and-sandbox` removes two different Codex barriers: the sandbox and approvals. We explain the full security model, read-only/workspace-write/full-access modes, rules, permission profiles, networking, CI, and how to design autonomous agents without handing them the whole machine.
A metacognition framework for LLMs proposes measuring correctness, conflict, experience, importance, and emotional load to choose between a fast path and multi-agent deliberation. The important idea is not giving AI consciousness, but making reasoning observable, measurable, and controllable.
A real migration from classic Application Insights URL Ping Tests to Standard Tests using Codex with GPT-5.6 Terra, Azure CLI, Azure MCP, and ARM, preserving alerts, secrets, end-to-end validation, costs, and rollback.
OpenAI plans to remove its models from Cursor on November 12, 2026 following the SpaceX acquisition. Beyond the corporate conflict, the episode exposes a key technical risk: depending on one model provider inside an agentic harness.
Technical debt makes software more expensive to change; operational debt makes it more expensive to keep running. We explain the difference, how both accumulate, how they reinforce each other, and why observability migrations such as URL Ping Tests → Standard Tests are a good example of both.
A viral 2023 tutorial showed how to put ChatGPT inside Jupyter by reusing the browser session. In 2026 the pattern is still valuable, but the implementation has changed: Jupyter AI, the official API, local models with Ollama, and executable notebooks turn the LLM into a true analysis agent.
A video by Gustavo Entrala shows email, calendar, presentations, spreadsheets, and automations controlled from ChatGPT. The key is not a spectacular demo: it is the architectural shift toward Work, apps, Cloud Browser, and Scheduled Tasks, where natural language becomes an interface for executing real work.
Capital de Tokens
