Home
Home Blog News
LinkedIn GitHub

Weekly News - #14


  • The machines are fine. I'm worried about us.

    True expertise is created through years of work, developing an instinct. While agentic tools can synthesize professional drafts in days, they lack memory of failure required to catch the subtle errors that only a seasoned human supervisor can detect.

  • Best Practices for Secure Error Handling in Go

    Errors in Go are just values, which means they'll leak your database schemas, file paths, and credentials straight into logs if you let them. The difference between secure and exposed systems comes down to data classification: structured logging with explicit field selection, redaction at the middleware layer, and strict separation between what gets recorded for debugging versus what crosses service boundaries.

  • How NASA Built Artemis II’s Fault-Tolerant Computer

    NASA has detailed the computing architecture for the Artemis II mission, moving away from the electromechanical fallbacks of the Apollo era toward a software-defined, ultra-resilient system. Given that hardware repair is impossible in deep space, the system is designed to handle radiation-induced faults without downtime.

© 2026 Maciej Szalewicz. All rights reserved.
LinkedIn GitHub