Is your AI feature a "Main Quest" helper or a high-friction detour?

Most AI feels like a "Pull" handle on a door that should have a Push-bar. I deconstructed an AI Alt-Text demo using the built-in AI APIs to meet users where they already go.

✅ Follow The Momentum: Why Push-bars beat Pull handles.
✅ The Latency Tax: Hide the seams or quality won't matter.
✅ Temporal Illusions: Why "instant" AI can decrease trust.

📖 https://exploractical.com/blog/2026/alt-text/
🧪 https://exploractical.com/demos/alt-text/

#WebAI #UX #BuiltInAI

Nico Martin (@nic_o_martin)

Transformers.js v4가 공개되었으며, 새로운 모델과 기능이 추가되고 성능도 더 빨라졌다고 소개한다. 자바스크립트 환경에서 최신 AI 모델을 더 쉽게 활용할 수 있는 중요한 업데이트로 보인다.

https://x.com/nic_o_martin/status/2038621948863713578

#transformersjs #javascript #webai #llm #opensource

🤷 Nico Martin (@nic_o_martin) on X

🤗 Transformers.js v4 is out! And I'm incredibly proud to be part of it! New models, new features, faster than ever 🔥 Check it out 👇

X (formerly Twitter)
New from my Web AI Lab: I’ve built an “article assistant” for my site that runs using local AI in the browser (via Chrome + Gemini Nano) when available — and falls back to a cloud model when it isn’t. I think local AI has huge implications for the #OpenWeb. ricmac.org/2026/03/19/a... #WebAI

Building an article assistant:...
Building an article assistant: local AI in the browser with cloud fallback - Richard MacManus

How I built an article assistant for my website that uses local AI (via Chrome + Gemini Nano) when available, and falls back to a cloud model when it isn’t.

Richard MacManus

New from my Web AI Lab: I’ve built an “article assistant” for my site that runs using local AI in the browser (via Chrome + Gemini Nano) when available — and falls back to a cloud model when it isn’t.

Before you dismiss this because it's AI, I think this has huge implications for the #OpenWeb. Instead of sending every interaction to BigTech clouds, users can increasingly run AI on their own device: better privacy, lower cost & more user control.

Check it out: https://ricmac.org/2026/03/19/article-assistant-local-ai-browser/ #WebAI

Building an article assistant: local AI in the browser with cloud fallback - Richard MacManus

How I built an article assistant for my website that uses local AI (via Chrome + Gemini Nano) when available, and falls back to a cloud model when it isn’t.

Richard MacManus

Nico Martin (@nic_o_martin)

브라우저에서 AI 실행을 쉽게 해주는 신규 기능 공개 소식입니다. Transformers.js의 AgentSkill을 출시해 웹앱 내에서 온디바이스 ML 모델을 구동하도록 코딩 에이전트에 필요한 정보를 제공하고, 브라우저 기반 AI 통합을 단축·간소화한다고 안내합니다.

https://x.com/nic_o_martin/status/2031773328441049270

#transformersjs #agentskill #webai #ondeviceai

🤷 Nico Martin (@nic_o_martin) on X

Getting started with AI in the browser has never been easier 🤗 Today we're releasing the Transformers.js #AgentSkill, your shortcut to running ML models in the browser. It provides all the information your coding agent needs to add on-device AI to your web app.

X (formerly Twitter)
I experimented with WebMCP on my personal website, exposing two tools an AI assistant can call directly from the browser: searching an article and subscribing to my newsletter. It’s a small prototype, but it hints at how websites are fast becoming AI-interactive surfaces. https://ricmac.org/2026/03/11/webmcp-ai-agents-interact-website/ #WebAI #WebMCP #WordPress
Implementing WebMCP: letting AI agents interact with my website - Richard MacManus

What happens when a website exposes tools to AI agents? To experiment, I implemented WebMCP on my personal site using two simple browser-side tools.

Richard MacManus

In most enterprise environments, third party cloud-based STT is a non-starter. That usually leaves us with mediocre OS defaults.

My colleague @paul 's "Utter" Chrome extension fixes this by keeping the intelligence loop local. It uses the WebSpeech and Prompt APIs to transcribe and polish your "train of thought" entirely on-device. 🛡️✨

A productivity unlock for AI agent workflows without the data leakage.

Give it a spin: https://chromewebstore.google.com/detail/utter/eejdhhjghkhkfejhbceeknhebhegijee

#WebAI #BuiltInAI

Utter - Chrome Web Store

Global hotkey for voice-to-text input using Speech Recognition

As part of my #WebAI explorations, I've built an AI chatbot for my personal website called Ask Ricmac. Under the hood, it runs on a Cloudflare Workers backend that uses Vectorize, D1 and Workers AI. During development, I also used the WordPress MCP Adapter and Claude Desktop. In this post I explain how these pieces fit together and the role each one plays. https://ricmac.org/2026/03/06/building-ask-ricmac-my-first-experiment-in-the-web-ai-stack/
Building Ask Ricmac: my first experiment in the Web AI stack - Richard MacManus

Over the past few months I’ve been exploring what I think of as the Web AI stack — the emerging intersection of artificial intelligence with the open web. As part of that exploration, I built a small experiment for my personal website: an AI chatbot called Ask Ricmac. Its purpose…

Richard MacManus
While I figure out what’s next in my career, I resolved to dive into a technology stack I’ve been deeply interested in for a while: Web AI. Only this time not just writing about these technologies, but building apps with them too. ricmac.org/2026/02/26/w... #WebAI

Becoming a Web AI Practitioner...
Becoming a Web AI Practitioner: A Map of the Emerging Stack - Richard MacManus

Earlier this month, I was laid off from my job as senior editor at The New Stack. While I figure out what’s next, I resolved to dive into a technology stack I’ve been deeply interested in for a while: Web AI. Only this time not just writing about these technologies,…

Richard MacManus
While I figure out what’s next in my career, I resolved to dive into a technology stack I’ve been deeply interested in for a while: Web AI. Only this time not just writing about these technologies, but building apps with them too. In this post, I’ve taken a stab at a high-level overview of the Web AI stack, which I'd love your feedback on. https://ricmac.org/2026/02/26/web-ai-stack/ #WebAI
Becoming a Web AI Practitioner: A Map of the Emerging Stack - Richard MacManus

Earlier this month, I was laid off from my job as senior editor at The New Stack. While I figure out what’s next, I resolved to dive into a technology stack I’ve been deeply interested in for a while: Web AI. Only this time not just writing about these technologies,…

Richard MacManus