THE DAILY SCRAPE

Your daily rundown on AI and what it means for the American Workforce

May 22, 2026

Friday, May 22, 2026

The enterprise AI adoption story is crystallizing around a familiar management challenge: how to run fast and slow simultaneously. SiliconANGLE’s reporting on the “two-speed” imperative captures what’s becoming the defining tension for large organizations—maintaining stable operations while experimenting with transformative technologies that could remake entire business models.

Successful AI transformation requires a dual-speed approach, distinguishing between workflows needing careful redesign and those benefiting from rapid experimentation. Organizations failing to adopt this multi-speed strategy risk hindering their AI adoption and losing competitiveness. SiliconANGLE

THERE’S A NEW WAY TO CREATE GOOGLE DOCS WITH YOUR VOICE

WSJ's Nicole Nguyen got early access to Docs Live, Google's voice-driven Docs feature that lets users ramble through ideas while Gemini structures a draft, pulls in relevant material from Drive, and adjusts on command. It joins Gmail Live and a Keep update in Google's broader push to make voice the input layer for Workspace — useful for spinning up checklists or first drafts, less so for finished prose. WSJ

TOO MUCH WORK? HAVE YOUR DIGITAL TWIN DO IT.

The WSJ profiles a small but growing set of executives running AI "digital twins" trained on their writing, speeches, and interviews — Reid Hoffman's "Reid AI" has delivered over 75 addresses since 2024, and Greif's CHRO twin has fielded interactions with nearly 3,300 employees. Open questions remain about ownership when the human leaves, hallucinations on judgment calls, and whether twins eventually replace rather than extend their originators. WSJ

HOW I CHOOSE WHICH CLOUDFLARE EMPLOYEES TO REPLACE WITH AI

Matthew Prince writes in the WSJ that Cloudflare laid off over 20% of its workforce while growing revenue more than 30%, targeting what Peter Drucker called "measurers" — middle managers, operations, internal audit, finance, compliance, marketing — rather than builders or sellers. Prince argues AI now measures organizations more continuously and precisely than humans can, and predicts the growth-with-layoffs pattern will become standard across the next year. WSJ

AI WILL FORCE LAWYERS TO CHANGE THEIR FEE STRUCTURE

In an FT AI Exchange interview, Harvey co-founder Winston Weinberg argues AI will push law firms from billable hours toward fixed fees for tasks like diligence, with partner advice priced separately at premium rates. He also expects in-house teams to refuse "crazy bills" for AI-doable work, forcing firms to rebuild the apprenticeship pipeline so junior associates learn high-judgment work rather than grinding through 1,000 reps of rote tasks. Financial Times

INDEED PARENT COMPANY TOUTS RECORD GROWTH DRIVEN BY AI

Recruit Holdings, Indeed’s parent company, reported record user growth in March despite anxieties surrounding AI’s potential disruption of job boards. The growth may even be driven by AI-related job postings or AI-powered features on the platform itself. HR Dive

SERVERLESS AI INFRASTRUCTURE STARTUP MODAL LABS SEALS $355M FUNDING ROUND

Modal Labs, a serverless AI infrastructure startup, has secured $355 million in funding to address the challenges of managing AI-generated code and providing sufficient computing power for AI applications. As the squeeze on compute resources tightens, Modal is an intermediary that helps firms close the gap. SiliconANGLE

CALIFORNIA TO EXPLORE POSSIBLE SUBSIDIES FOR COMPANIES THAT DON’T REPLACE WORKERS WITH AI

Governor Gavin Newsom has issued an executive order directing state officials to explore overhauling labor policies in anticipation of potential job displacement due to AI. The state is considering measures, potentially including subsidies, for companies that choose not to replace human workers with AI. Remember: what starts in California doesn’t stay there. SiliconANGLE

WISETECH BEGINS REDUNDANCIES – BUT OMITS ‘AI’ FROM EMAILS TO CHINESE EMPLOYEES· WORKERS SAY

WiseTech, an ASX-listed company, has begun laying off nearly 30% of its global workforce, citing advancements in artificial intelligence. However, the company reportedly omitted the term “AI” in redundancy emails sent to employees in China. If you don’t say the word “AI”, does that make displacement go away? Guardian Tech

OPENAI MAKES BREAKTHROUGH ON 80-YEAR-OLD MATHS PROBLEM

OpenAI claims to have made a breakthrough in AI reasoning by successfully tackling the 80-year-old planar unit distance problem posed by Paul Erdős. Math is the most basic input to most technological advancement; this evolution will end up driving innovation of all types. Guardian Tech

INSIDE ANTHROPIC’S PUSH TOWARD PROFITABILITY—AND THE 1 BIG HURDLE AHEAD

Anthropic, an AI company, is reportedly aiming for profitability, projecting $10.9 billion in revenue for the second quarter. Anthropic is on the leading edge of several exponential AI curves, from model development, to tool development, to deployment. Not surprisingly, that is now leading to profit. Inc.

SPOTIFY STUDIO’S AI AGENT CREATES A DAILY PODCAST JUST FOR YOU

Spotify Labs has launched “Studio,” an AI app that creates personalized daily briefings, podcasts, and playlists. The AI uses a user’s Spotify listening history and connected apps like email and calendars to generate tailored content. The Verge - AI

Videos & Podcasts

▶ Google I/O 2026, Karpathy Joins Anthropic, and Cerebras’ $95B IPO | EP #256— Moonshots (Peter Diamandis)

▶ How ICE Is Hurting Local Economies— Bloomberg Television

▶ How to Lead People Through AI Change: Questions to Ask from a Transformation Expert— World Economic Forum

▶ AI is now building itself?— Alex Kantrowitz

Quick Scan

Do AI Risks Require Extraordinary Government Intervention? — AI Snake Oil

Zscaler acquires Symmetry Systems to extend its AI agent security capabilities — SiliconANGLE

Hark raises $700M+ to build ‘personalized intelligence’ devices — SiliconANGLE

Bugcrowd launches reinforcement learning environments to train AI on real software vulnerabilities — SiliconANGLE

AI interoperability and layered trust emerge as the real unlocks for enterprise scale — SiliconANGLE

Impetus shifts Leap AI toward enterprise context engineering — SiliconANGLE