🟣 TOOOLS #24: Doop, Causal, Pryzm, Doodle Icons, Neurath X

   TOOOL OF THE WEEK

   ALSO ON THE RADAR

Causal

An infinite canvas for moodboards, notes, files and links on the Mac, with an AI that reads the whole board and places new ideas where they belong.

Pryzm

A browser studio for backgrounds and animated loops nobody else has, blended from gradients, glass and grain, exported as PNG, video or a live embed for your site.

Doodle Icons

A growing set of hand-drawn icons that sketch themselves in stroke by stroke and never sit still, exported as animated SVG and open source.

Neurath X

A geometric sans from Studio René Bieder in proportional and monospaced cuts, reduced and optically refined for editorial, branding and interface work.

   FROM THIS WEEK'S SPONSOR

   A FEW READS & LISTENS

UX COLLECTIVE  📖 ARTICLE

When the canvas starts acting, who's really in control?

Aurélie Radom on what changes once the canvas can act on its own: the job shifts from specifying every interaction to deciding what an agent may do unasked, what needs a confirmation, and what always waits for a person. Her sharpest point is that attention becomes the material, and the design system turns into standing instructions for the agent.

DESIGN BUDDIES  📖 ARTICLE

Hiring in the Age of AI in Design

Carl Wheatley, a design recruiter with a decade at Meta, pulls the 2026 hiring data into one place: 82 percent of hiring managers need as many designers or more, 73 percent want AI fluency, and junior roles keep shrinking as the repetitive work people used to learn on disappears. Practical on the portfolio side too, including why being open about where AI helped now reads as good judgment.

DIVE CLUB  🎧 PODCAST

The story of Evil Rabbit and designing Vercel

Vercel's founding designer on a first design job at twelve in Argentina, building one of the most copied brand identities in software, and why the team moved from producing assets to building its own tools. The closing stretch on Lifeline, a personal timeline he built with an agent in a few days and then open-sourced, shows where that shift ends up.

LENNY'S NEWSLETTER  📖 ARTICLE

How to turn your AI into a world-class designer

Anshu Chimala, formerly designing at Apple, on why AI output looks generic by default (a model picks the safest choice at every step) and how to push it off that path: random seed strings for real variety, a second AI as a critic scoring against a quality bar, generated images and motion for personality, then cutting every glow and extra label. The back half sits behind Lenny's paywall, but the method is all in the open part.

   UPDATES & NEWS

Google Pics puts image generation and editing straight into Docs and Slides: isolate one object and change it, or rewrite the text inside a picture, without leaving the document

Adobe for Slack puts 70+ tools from Firefly, Express, Photoshop, Premiere and Acrobat behind a plain-language request to Slackbot, so a thread's context becomes an image, a video or a PDF without leaving the channel, on Business+ and Enterprise+ plans

OpenAI releases GPT-6 Astra: it works your apps on its own, lays out slides with an actual narrative, and through ChatGPT's Sites turns one prompt into a hosted website or web app

Anthropic releases Claude Fable 5.1 and Claude Mythos 5.1, one model at two safeguard levels: Fable open to everyone on every platform from day one, Mythos reserved for vetted security and life-science teams

The Twitter bird is back as Twitter.now, relaunched by Twitter's former trademark lawyer on the claim that X abandoned the brand, phoenix feathers on the logo and a fail whale on the error page

   ONE MORE THING ...

SECOND ACT

"Nineteen years of Awwwards, then the CEO builds his own Site of the Day."