Bob Z
Subscribe
Bob Zhu - Bob Z
Bob Zhu

Bob Zhu

Ink sketch of an open notebook with edited prose on one page, a code window on the other, and a plum pencil between them.
Essays

Astra for Writing, Fable for Coding

Astra feels better for writing. Fable feels better for coding. Live speech adds another use case: talking through ideas and steering a draft as it develops.
Hand-drawn editorial diagram of a prompt entering a layered package of context, permissions, tools, checks, and workflow.
Essays

The Prompt Is Becoming a Package

Reusable AI work is moving out of the clever prompt and into packages of context, tools, instructions, permissions, examples, and workflows.
Hand-drawn editorial illustration of expert constraint cards steering an execution machine through a maze.
Essays

AI Makes Expertise More Valuable Before It Makes It Less Valuable

Current agents lower implementation barriers while rewarding the domain knowledge needed to specify work, catch edge cases, and recover from failure.
Essays

The Useful Metric Is Cost Per Finished Task

The July 2026 model releases are argued in cost, tokens, and tool calls per task rather than leaderboard position. The harder problem is that most workflows can't say what finished means.
Hand-drawn editorial illustration of a microphone routing one spoken instruction into browser, coding, desktop, and phone tasks.
Essays

GPT-Live Is the Voice Assistant I Actually Wanted

ChatGPT Voice turns speech into a control layer for real work: browse the web, steer Codex tasks, share screen context, and keep the same desktop environment moving from your phone.
Hand-drawn editorial illustration contrasting a machine that produces loose pages from tokens with a connected system of notes, scripts, checks, and documentation.
Essays

Build Something That Stays

The useful difference is not who uses the most AI. It is whether each task leaves behind something that makes the next one cheaper, safer, or more consistent.
A hand-drawn desk-garden scene where a software interface grows from an open notebook, with pruning shears and connected app windows around it.
Essays

Organic Software Is Not Vibe Coding

Organic software is expert-led, AI-assisted software that grows through real use. It is dynamic, but not loose: the builder keeps changing the app as the work reveals itself, while using judgement, tests, and feedback to stop the system from becoming a mess.
Hand-drawn editorial illustration of an open notebook branching into several AI model routing lanes, with abstract model cards, cached-context notes, and a stopwatch loop.
Essays

The Interesting Part of GPT-5.6 Is the Routing Layer

GPT-5.6 Sol, Terra, Luna, and the Cerebras launch point at a more practical question: which lane should this work run in, how fast should it move, and when should a human stay close?
Sketch illustration of a person at a desk looking at an eight-step AI proficiency ladder that rises from chat to orchestration.
Essays

The AI Proficiency Ladder

Every's eight-level AI adoption map is useful because it measures delegation, trust, context, and verification, not personal intelligence. This version expands the ladder and pairs it with a self-assessment that checks how people actually work with AI.
Hand-drawn editorial illustration of a notebook, stopwatch, and small task loops branching from one human decision point.
Essays

The Future Feels Like 1,000 Tokens a Second

Cerebras is interesting beyond its IPO because near-1,000-token-per-second inference changes the shape of the work: draft, check, repair, compare, and return before the human loses the thread.
Hand-drawn editorial illustration of an open AI floor rising toward a proprietary AI ceiling, with model blocks, agent loops, and benchmark markers.
Essays

The Floor and Ceiling of AI

The useful question is how much AI capability would still remain if the proprietary frontier disappeared tomorrow.
Minimal hand-drawn illustration of an open notebook with a symbolic clock gauge and small markers for video, support, code, and medical admin functions.
Essays

The AI Function Clock

The common claim is that AI cannot replace your job. That may be true. But jobs are bundles of functions, and the better question is which functions are already moving closer to reliable AI handoff.
Hand-drawn editorial illustration showing a central rule document under pressure from opposing loophole and overreach forces, with review and precedent hovering around it.
Essays

Loophole Treats Ethics Like Adversarial Testing

Loophole is not interesting because it solves ethics. It is interesting because it treats moral principles like something you can draft, attack, patch, and escalate until the real conflicts in your values finally surface.
Hand-drawn editorial illustration showing a rigid dashboard evolving into a chat bar and then into a generated multi-panel interface.
Essays

The SaaS Homepage Is Becoming an Agent

Attio, Linear, and PostHog are converging on the same product bet: one fixed dashboard cannot serve every user or every intent, so the SaaS homepage is shifting toward an agentic entry point. Chat is the router. Generative UI is what comes next.
Sketch illustration of an open notebook surrounded by five agent communication patterns: paired review, handoff, shared thread, swarm routing, and an external capability card.
Essays

How Agents Actually Talk to Each Other

The next real step for agents is genuine multi-agent runtime across different harnesses: paired review loops, handoffs, shared threads, and open protocols such as A2A.
Sketch editorial illustration of a phone splitting into a local laptop workflow on one side and cloud agent workspaces on the other.
Essays

Mobile AI Coding Is Splitting Into Local Supervision and Cloud Autonomy

As of March 24, 2026, mobile AI coding is splitting around where the work runs: a phone steering your own machine, or a phone watching cloud agents work somewhere else.
Sketch illustration of an open notebook turning into five stages of autonomous Claude Code workflows, ending in a measured results chart.
Essays

Five Levels of Running Claude Code More Autonomously

Claude Code autonomy is not one trick. It starts with permission friction and context hygiene, then moves through subagents and loop hooks, and becomes genuinely useful when the system can measure whether the last iteration actually improved the work.
Sketch illustration of a central notebook page turning into a layered memory system with cards, media inputs, and timeline markers.
Essays

Agent Memory Is Being Built Around the Model

The newest memory systems do not make language models inherently stateful. They build recall, updates, and temporal reasoning around the model instead. Gemini Embedding 2 is the first mass-scale vector embedding that is multimodal.