Weekly Web Harvest for 2026-04-12

Out of the box, the model will natively interpret a transcript and determine how your words should be delivered. Simple transcripts without any additional prompting sound natural. But Gemini TTS also comes with tools you can use to steer it. The purpose of this guide is to offer fundamental direction and spark ideas when developing audio experiences. We’ll start with Tags for quick inline control, and then explore advanced Prompting structures for full performance direction.

Source: Text-to-speech generation (TTS)  |  Gemini API  |  Google AI for Developers

I think we will see some people employed (though perhaps not explicitly) as meat shields: people who are accountable for ML systems under their supervision. The accountability may be purely internal, as when Meta hires human beings to review the decisions of automated moderation systems. It may be external, as when lawyers are penalized for submitting LLM lies to the court. It may involve formalized responsibility, like a Data Protection Officer. It may be convenient for a company to have third-party subcontractors, like Buscaglia, who can be thrown under the bus when the system as a whole misbehaves. Perhaps drivers whose mostly-automated cars crash will be held responsible in the same way—Madeline Clare Elish calls this concept a moral crumple zone.

Source: The Future of Everything is Lies, I Guess: New Jobs

A cognitive scaffolding intervention (partnership training that reframed AI as a thought partner) was associated with higher individual document quality at the top of the distribution.

Source: [2604.08678] Scaffolding Human-AI Collaboration: A Field Experiment on Behavioral Protocols and Cognitive Reframing

h/t Downes

The findings reveal a counterintuitive paradox: pedagogical partnerships simultaneously activate cognitive vigilance (critical evaluation) and cognitive offloading (strategic delegation). Contrary to traditional assumptions that delegation diminishes learning depth, our results indicate that strategic cognitive offloading—when exceeding specific thresholds—liberates mental resources for higher-order reflection, thereby enhancing transformative learning experiences

Source: Pedagogical partnerships with generative AI in higher education: how dual cognitive pathways paradoxically enable transformative learning | International Journal of Educational Technology in Higher Education | Springer Nature Link

So I sat next to him and we started examining the UI changes in detail. And Cabel was zooming in and out instantaneously. I was like, “Whoa, how are you doing that?” And Cabel was like, “You don’t know about the Accessibility Zoom gesture? Here, let me show you!” And my mind was blown. Cabel emphasized the importance of going into the “Advanced…” dialog to turn off “Smooth images”, and I agree. This is a fantastic feature, but Apple has the default setting wrong for smoothing (a.k.a. blurring) the zoomed image. I honestly can’t imagine why anyone would want the zoomed image blurred.

Source: Daring Fireball: MacOS Tip: Enable the Zoom ‘Peek’ Gesture

Must be able to ignore the most foul stench coming off my open leg wound, and not ask any pointed questions about the double-edged fate that befell my most recent wife.

Source: Henry VIII’s Dating App Profiles – McSweeney’s Internet Tendency

Ella Freire

Source: Pan American Luggage Labels – Ella Freire

Allbirds, the San Francisco footwear company that recently agreed to sell its core shoe business, said Wednesday it plans to reinvent itself as an artificial intelligence company — a shift that sent its stock surging in early trading.

Source: SF shoe brand Allbirds pivots to AI, shares surge

Shares in a tiny US soft drinks firm, Long Island Iced Tea Corp, have quadrupled in value in just one day after it changed its name to Long Blockchain Corporation, the latest move by a string of obscure companies rebranding themselves in an attempt to ride the bitcoin bubble.

Source: Tiny US soft drinks firm changes name to cash in on bitcoin mania | Bitcoin | The Guardian

Despite their ubiquity, wireless earbuds remain audio-centric due to size and power constraints. We present VueBuds, the first camera-integrated wireless earbuds for egocentric vision, capable of operating within stringent power and form-factor limits. Each VueBud embeds a camera into a Sony WF-1000XM3 to stream visual data over Bluetooth to a host device for on-device vision language model (VLM) processing. We show analytically and empirically that while each camera’s field of view is partially occluded by the face, the combined binocular perspective provides comprehensive forward coverage. By integrating VueBuds with VLMs, we build an end-to-end system for real-time scene understanding, translation, visual reasoning, and text reading; all from low-resolution monochrome cameras drawing under 5mW through on-demand activation.

Source: VueBuds: Visual Intelligence with Wireless Earbuds

Suffer like the artist you are. It ain’t easy, but if it were easy, it wouldn’t be worth doing.

Source: Opinion | Don’t Use A.I. to Do This – The New York Times

— The piece that I struggle with is that many, many people are just doing a thing for money (even art) and that argument holds no weight there. Maybe that’s the final stage of capitalism? Get by with whatever you can so you get paid as you wait for the thing you’re using to get by to replace you.

h/t Sara P.