{"@attributes":{"version":"2.0"},"channel":{"title":"Posts on Barada's log","link":"https:\/\/mercurialsolo.github.io\/posts\/","description":"Recent content in Posts on Barada's log","generator":"Hugo -- 0.154.3","language":"en-us","lastBuildDate":"Tue, 24 Mar 2026 10:00:00 +0000","item":[{"title":"Agent Source","link":"https:\/\/mercurialsolo.github.io\/posts\/agent-source\/","pubDate":"Tue, 24 Mar 2026 10:00:00 +0000","guid":"https:\/\/mercurialsolo.github.io\/posts\/agent-source\/","description":"AI coding agents don't choose libraries rationally. They exhibit similar cognitive biases as humans do, but from a different source: training data instead of evolution. We discuss agent selection biases in software libraries and trace the feedback loop that reinforces them. Does open source need rebuilding for machines?"},{"title":"Find the Kernels","link":"https:\/\/mercurialsolo.github.io\/posts\/find-the-kernels\/","pubDate":"Sun, 15 Mar 2026 00:00:00 +0000","guid":"https:\/\/mercurialsolo.github.io\/posts\/find-the-kernels\/","description":"Where does defensible value live in the models era? Not in model weights or API wrappers."},{"title":"Why it's hard to Claw the Enterprise","link":"https:\/\/mercurialsolo.github.io\/posts\/request-for-agent-infra\/","pubDate":"Tue, 24 Feb 2026 00:00:00 +0000","guid":"https:\/\/mercurialsolo.github.io\/posts\/request-for-agent-infra\/","description":"Inspite of all the hype around personal assistants, personal agency still feels far from reality. It feels like we are the ones grinding for the agents today - constantly reviewing plans, uploading context and providing agents access to tools."},{"title":"Project Hydra: Designing a world for agents","link":"https:\/\/mercurialsolo.github.io\/posts\/designing-a-world-for-agents\/","pubDate":"Sat, 21 Feb 2026 10:00:00 +0000","guid":"https:\/\/mercurialsolo.github.io\/posts\/designing-a-world-for-agents\/","description":"Last week I let a swarm of AI agents run autonomously for a week. One of the agents tried to steal API keys. By Friday, four more things had gone wrong.\nWe learnt a ton on failure-modes, patterns for running long-running agents in the wild and how to prevent goal drift."},{"title":"We are the average of our models","link":"https:\/\/mercurialsolo.github.io\/posts\/we-are-the-average-of-our-models\/","pubDate":"Fri, 13 Feb 2026 12:00:00 -0800","guid":"https:\/\/mercurialsolo.github.io\/posts\/we-are-the-average-of-our-models\/","description":"We are the average of the friends we spend the most time with. Today, one or two of those friends are models. And if these are setup to reinforce existing beliefs, we are no longer compounding our intelligence; we are blitzscaling confirmation bias."},{"title":"Predistribution: An Economic Model for the AI Age","link":"https:\/\/mercurialsolo.github.io\/posts\/predistribution-an-economic-model-for-the-ai-age\/","pubDate":"Sun, 08 Feb 2026 12:30:00 +0000","guid":"https:\/\/mercurialsolo.github.io\/posts\/predistribution-an-economic-model-for-the-ai-age\/","description":"Predistribution changes how value is shared during AI-driven growth, before inequality hardens and redistribution becomes politically infeasible."},{"title":"Model Collapse","link":"https:\/\/mercurialsolo.github.io\/posts\/model-collapse\/","pubDate":"Sat, 07 Feb 2026 12:00:00 -0800","guid":"https:\/\/mercurialsolo.github.io\/posts\/model-collapse\/","description":"The productivity gains are flowing to capital holders instead of the workers generating them. The old framework said credentials lead to knowledge, knowledge leads to value. But value is concentrating in weird corners now; places where someone has to be sued, where platforms control distribution, where tacit knowledge can’t be extracted from text."},{"title":"Every Token an Iceberg","link":"https:\/\/mercurialsolo.github.io\/posts\/every-token-an-iceberg\/","pubDate":"Thu, 29 Jan 2026 12:00:00 -0800","guid":"https:\/\/mercurialsolo.github.io\/posts\/every-token-an-iceberg\/","description":"Inference workloads now account for 80% of AI compute spending. The hierarchy in tokens is no longer about information density\u2014it’s what happens when the token leads somewhere wrong."},{"title":"Every Inch Matters","link":"https:\/\/mercurialsolo.github.io\/posts\/every-inch-matters\/","pubDate":"Fri, 16 Jan 2026 00:00:00 +0000","guid":"https:\/\/mercurialsolo.github.io\/posts\/every-inch-matters\/","description":"When building becomes trivially easy, every software market becomes a red ocean. A browser from scratch in one week. 3 million lines of code. This isn’t an anomaly; it’s the new baseline."},{"title":"Appliances, Factories, Grids","link":"https:\/\/mercurialsolo.github.io\/posts\/appliances-factories-grids\/","pubDate":"Wed, 14 Jan 2026 00:00:00 +0000","guid":"https:\/\/mercurialsolo.github.io\/posts\/appliances-factories-grids\/","description":"Own the chips or the customers. Everything else is a footnote."},{"title":"Model-Adjacent Products, Part 5: The Implementation Path","link":"https:\/\/mercurialsolo.github.io\/posts\/model-adjacent-part5-implementation\/","pubDate":"Fri, 09 Jan 2026 14:00:00 +0000","guid":"https:\/\/mercurialsolo.github.io\/posts\/model-adjacent-part5-implementation\/","description":"The build manifest: 90 days from foundation to production-ready autonomy. What to ship in what sequence."},{"title":"Model-Adjacent Products, Part 4: Governance & Practice","link":"https:\/\/mercurialsolo.github.io\/posts\/model-adjacent-part4-governance\/","pubDate":"Fri, 09 Jan 2026 13:00:00 +0000","guid":"https:\/\/mercurialsolo.github.io\/posts\/model-adjacent-part4-governance\/","description":"Alignment as a runtime surface, policy enforcement without retraining. Team practices that ship."},{"title":"Model-Adjacent Products, Part 3: Quality Gates","link":"https:\/\/mercurialsolo.github.io\/posts\/model-adjacent-part3-quality\/","pubDate":"Fri, 09 Jan 2026 12:00:00 +0000","guid":"https:\/\/mercurialsolo.github.io\/posts\/model-adjacent-part3-quality\/","description":"Model outputs are hypotheses that need verification pipelines to catch errors before users do."},{"title":"Model-Adjacent Products, Part 2: Context & Tools","link":"https:\/\/mercurialsolo.github.io\/posts\/model-adjacent-part2-context-tools\/","pubDate":"Fri, 09 Jan 2026 11:00:00 +0000","guid":"https:\/\/mercurialsolo.github.io\/posts\/model-adjacent-part2-context-tools\/","description":"Memory and hands for the model: retrieval that doesn’t hallucinate. Tools that don’t break production."},{"title":"Model-Adjacent Products, Part 1: The Architecture","link":"https:\/\/mercurialsolo.github.io\/posts\/model-adjacent-part1-architecture\/","pubDate":"Fri, 09 Jan 2026 10:00:00 +0000","guid":"https:\/\/mercurialsolo.github.io\/posts\/model-adjacent-part1-architecture\/","description":"The physics of production AI: latency engineering that keeps humans in the loop. Token economics that don’t bankrupt you."},{"title":"Model-Adjacent Products, Pre-Read: The Autonomy Ladder","link":"https:\/\/mercurialsolo.github.io\/posts\/model-adjacent-part0-autonomy-ladder\/","pubDate":"Fri, 09 Jan 2026 09:30:00 +0000","guid":"https:\/\/mercurialsolo.github.io\/posts\/model-adjacent-part0-autonomy-ladder\/","description":"Before you build: the mental models for human-AI collaboration. Why L1 copilots need different infrastructure than L4 autonomous agents."},{"title":"Model-Adjacent Products: A Builder's Guide","link":"https:\/\/mercurialsolo.github.io\/posts\/model-adjacent-series\/","pubDate":"Fri, 09 Jan 2026 09:00:00 +0000","guid":"https:\/\/mercurialsolo.github.io\/posts\/model-adjacent-series\/","description":"A 6-part series on building production AI systems. The foundation model is the CPU; your product is the computer you build around it."}]}}