Article URL: https://filipovski.net/2026/09/16/backups-arent-simple.html Comments URL: https://news.ycombinator.com/item?id=49732513 Points: 10 # Comments: 0
This is an AI-generated summary. ShortSingh links to the original source for the complete article.
Most large organizations today operate hybrid delivery models that blend Agile sprints with traditional governance gates, even when they do not formally acknowledge doing so. The Agile Manifesto, authored by 17 software developers in 2001, was designed for small teams rather than complex, regulated enterprises with multi-layered oversight structures. Core enterprise controls around governance, planning, and prioritization remain built for predictability, which has kept pure Agile confined to isolated pockets rather than embedded across strategy. Critics argue the 'Agile Industrial Complex' — consultants and certification bodies focused on process over culture — has accelerated what Martin Fowler calls 'faux Agile,' producing organizations that run retrospectives but cannot adjust a release date without board approval. Hybrid frameworks like Water-Scrum-Fall are widely dismissed in coaching circles, yet they persist because they address real structural needs that pure Agile ideology has consistently underestimated.
A developer has launched The World Wall, a website that lets anyone leave a short message, name, or drawn signature as a permanent entry on a shared digital wall. Each new visitor receives a unique sequential number, with the wall growing continuously as more people participate. The platform requires no account, email, or profile — users simply submit their mark and move on. So far, people from 106 countries have already left their entries on the wall. The project was inspired by the simple idea of giving everyone on the internet a place to say they existed.
Cloudflare has launched a 'Disallow AI Training' setting that allows website owners to prevent their content from being used to train AI models without affecting search engine indexing. The feature improves on the older 'Block AI Bots' control, which could inadvertently reduce a site's search visibility by treating all AI-related crawlers the same way. Cloudflare now classifies crawlers into three categories — Search, Training, and Agent — enabling more targeted access preferences. The no-training directive is published via robots.txt through Bot Preference Sync, and major crawlers from Apple, Google, and Microsoft have committed to honoring it. Training crawlers from Amazon, Anthropic, Meta, and OpenAI are expected to be blocked under this preference, while Bingbot compliance is anticipated by early 2027.
A new technical guide demonstrates how individuals can fine-tune Meta's Llama-3 language model entirely on an Apple Silicon Mac to privately analyze a decade of personal health data. The workflow uses Apple's MLX framework and a technique called LoRA (Low-Rank Adaptation), which trains only small adapter layers rather than the full model, making it feasible on consumer hardware with 32GB or more of unified memory. Health records such as PDFs are first converted into structured JSON datasets, then used to teach the model to identify trends in personal biomarkers like cholesterol or glucose over time. Because all processing happens locally, no sensitive medical data is transmitted to external cloud servers. The approach requires Python 3.11 and the mlx-lm library, and is presented as an accessible weekend project for privacy-conscious users with compatible Apple hardware.
Discussion (0)
Log in to join the discussion and vote.
Log in