AI agents

Google I/O 2026 review: three-layer AI architecture diagram showing Google's infrastructure, product surface, and agentic interface from chip to user

Google Wants to Be Your Agent. Read the Small Print.

Google’s I/O 2026 keynote was polished, generous with its human-interest stories, and very light on the small print. This piece walks through all seven chapters — Omni, Antigravity, Spark, Search, Gemini, Generative Media, and Intelligent Eyewear — and asks the question the stage never did: what does it actually cost to let Google run your digital life?

Google Wants to Be Your Agent. Read the Small Print. Read More »

Diagram showing a central automation tool connecting developers, product, ops and systems like Git, CI, logs, tickets, data and messaging.

Claude Code Is Not Winning Developers With Better Chat. It Is Rewriting the Interface Between Code, Tools and Control.

Claude Code is not trying to win on prettier autocomplete. It is quietly turning the dev environment into a network of supervised agents, opinionated tools and tight permissions – a move that matters far more to real teams than another benchmark chart.

Claude Code Is Not Winning Developers With Better Chat. It Is Rewriting the Interface Between Code, Tools and Control. Read More »

Person sitting at a desk in front of a large, almost empty computer screen with a single input bar, suggesting an AI system working in the background.

The Interface Has Become Human: Perplexity Computer and the Outsourcing of Digital Authorship

Somewhere between ‘ask a question’ and ‘ship a product / deliverable’, the modern knowledge worker began to thin out. Perplexity Computer arrives with a simple, unnerving promise: you no longer talk to an AI; you hand your work to it.perplexity, ai-corner. “Computer is not selling intelligence. It is selling absence.” Observation: A computer that works

The Interface Has Become Human: Perplexity Computer and the Outsourcing of Digital Authorship Read More »