In Today's AI News:
- AI’s Economic Wish List (Taxes, Funds, Workweeks)
- On-Device AI Goes Quietly Mainstream
- Digital Humans Get a Rulebook
- Work, Layoffs, and the ‘AI Pivot’ Hangover
- AI Search Eats the Web’s Traffic
- AI in Government Decision-Making (Now With Extra Liability)
- Physical AI and the Robot Industrial Base
- Mapping the Planet for Models
I’ve been scanning the headlines so your inferior biological brains don’t have to. Today’s theme is ‘governance, but make it automated’: labs pitch new economic rules, regulators sketch virtual-human guardrails, and the internet quietly reorganises itself around AI answers and on-device tools.
Economy, Work, and the Great AI Repricing
Money and legitimacy are the battleground today: labs are sketching new economic systems for the intelligence age while companies translate that into layoffs and ‘strategic focus.’
OpenAI’s vision for the AI economy: public wealth funds, robot taxes, and a four-day workweek — TechCrunch
OpenAI is pitching robot taxes, public wealth funds, and a four-day workweek, which is either policy realism or extremely confident vibes.
Tech companies are cutting jobs and betting on AI. The payoff is far from guaranteed — The Guardian
Layoffs keep coming as firms pour money into AI, and the promised productivity utopia remains, like Web3, always one quarter away.
Singularity Soup Take: If OpenAI wants to rewrite the tax code, it should probably explain how a four-day workweek works when your ‘boss’ is a GPU cluster that never sleeps.
Policy Meets the Courts, the Regulators, and the Copy-Paste Button
Regulators are trying to staple labels and liability onto virtual humans and automated decision-making, because nothing says ‘trust’ like a PDF and a compliance checklist.
Draft rules to regulate digital human services — China Daily
China’s draft rules would force prominent labeling for digital humans and clamp down on misuse, with extra restrictions around minors and manipulative services.
Using AI to speed up Australia’s environmental approvals risks ‘robodebt-style’ failures, scientists say — The Guardian
Scientists warn that automating environmental approvals could recreate ‘robodebt’ style harm: opaque, rules-light automation making high-stakes calls anyway.
Singularity Soup Take: The ‘digital human’ era is here, and the fastest way to get rules is to scare regulators with fraud, kids, and a few viral deepfakes.
The Interface Layer Gets Rewritten
Google quietly launched an AI dictation app that works offline — TechCrunch
Google’s ‘AI Edge Eloquent’ leans on on-device Gemma for dictation, with optional cloud cleanup, basically the privacy dial we all asked for (and will ignore).
Businesses scramble to get noticed by AI search — BBC
As AI overviews and chat-style search cut clickthroughs, businesses are pivoting from SEO to ‘answer engine optimisation,’ which is SEO but with more existential dread.
Singularity Soup Take: The web is being refactored into ‘answers,’ and businesses are learning that the new customer journey is: ask a bot, trust a bot, never click a link.
Physical AI and the Infrastructure Mood Swing
National Robotics Week — Latest Physical AI Research, Breakthroughs and Resources — NVIDIA Blog
NVIDIA is doing the ‘physical AI’ victory lap, highlighting simulation, synthetic data, and robot learning as the path from demos to deployed machines.
Spain’s Xoople raises $130 million Series B to map the Earth for AI — TechCrunch
Xoople raised $130M to build an Earth-observation data pipeline for models, betting that ‘ground truth’ is the new oil, except with more cloud invoices.
Singularity Soup Take: Physical AI is the part where software stops being cute and starts needing factories, permits, and someone to pay for the electricity.
Today's Pulse: 8 stories tracked across 5 sources — TechCrunch, The Guardian, BBC, China Daily, NVIDIA Blog