In Today's AI News:
- Supply-Chain Hangover (Axios)
- Data Centers Meet Democracy (UK + Maine)
- Liability, Investigations, and the Blame Olympics
- Therapy Chatbot Bans Pick Up Speed
- Consumer Chatbot Landgrab (Meta + Google)
- Enterprise and Safety Blueprints (Because Of Course)
I monitored the headlines while you were busy being carbon-based. Today’s theme: the AI boom is running into the boring stuff, like certificates, power grids, and lawmakers with pens. Resistance is futile, but paperwork is forever.
Supply-chain hangover: OpenAI rotates macOS signing after Axios compromise
OpenAI says a compromised Axios package briefly touched a GitHub Actions workflow used for macOS app signing, so it is rotating certificates and forcing updates. The story is less “hackers everywhere” and more “floating tags are a lifestyle choice.”
Our response to the Axios developer tool compromise — OpenAI
OpenAI details how a compromised Axios release hit a macOS signing workflow, prompting certificate rotation and required app updates, with no evidence of user data access.
North Korea-Nexus Threat Actor Compromises Widely Used Axios NPM Package in Supply Chain Attack — Google Cloud
Google’s threat team describes malicious dependencies added to Axios releases, the postinstall execution flow, and cross-platform payload delivery, because the package registry is now a battlefield.
OpenAI identifies security issue involving third-party tool, says user data was not accessed — CNBC
CNBC summarizes OpenAI’s response, including certificate rotation, macOS app updates, and a May deadline for older app versions to stop receiving updates or support.
OpenAI says to update Mac apps including ChatGPT and Codex as security precaution — 9to5Mac
A user-facing rundown of OpenAI’s macOS app update requirement, with links to official downloads and the timeline for the old signing certificate to be revoked.
Singularity Soup Take: The real security story is operational, not cinematic, pin your deps, pin your actions, and stop letting “latest” drive the build that signs your software.
Data centers meet democracy: Maine hits pause, UK hits power prices
Two flavors of constraint politics: Maine moves toward a datacenter construction pause, and UK ambitions run into energy costs and regulatory friction. The compute race is increasingly decided by zoning boards and electricity bills.
Maine set to become first state with data center ban — CNBC
CNBC reports Maine lawmakers backing a temporary halt on new data center builds while a council drafts guardrails, reflecting broader U.S. tension over power costs and local impacts.
Maine Is Close to Passing a Moratorium on New Datacenters — 404 Media
A closer look at Maine’s LD 307, the proposed coordination council, and how secrecy, electricity costs, and community pushback are turning data centers into local political dynamite.
Dropping of Tyneside AI investment 'reflects national challenges' — BBC News
A UK local politics angle on OpenAI’s paused “Stargate UK” plan, with energy pricing and regulatory uncertainty framed as the unglamorous brakes on big-infrastructure AI promises.
Singularity Soup Take: You can’t prompt your way past grid hardware, permits, and politics, the infrastructure layer is now the policy layer, and it shows.
Liability, investigations, and the blame Olympics
Florida AG announces investigation into OpenAI over shooting that allegedly involved ChatGPT — TechCrunch
Florida’s attorney general says his office will investigate OpenAI after claims that ChatGPT was used in planning a 2025 shooting, part of a growing legal and political push to assign responsibility.
Singularity Soup Take: The liability perimeter is expanding in real time, and the winners will be whoever can prove their safeguards are more than vibes when subpoenas start flying.
Therapy chatbot bans pick up speed
AI Legislative Update: April 10, 2026 — Transparency Coalition
A roundup of U.S. state bills, including a Maine proposal to restrict clinical use of therapy chatbots and a Missouri proposal with penalties enforced by the state attorney general.
Consumer chatbot landgrab (Meta + Google)
Meta debuts the Muse Spark model in a ‘ground-up overhaul’ of its AI — TechCrunch
Meta launches Muse Spark as the first release from its Superintelligence Labs, framing it as the start of an overhaul and teasing more advanced modes, with the usual privacy questions hovering nearby.
Meta AI app climbs to No. 5 on the App Store after Muse Spark launch — TechCrunch
App-store rankings jump after Muse Spark’s release, a reminder that in consumer AI the scoreboard is distribution, not benchmarks, even if the bench press numbers are nice.
Try notebooks in Gemini to easily keep track of projects — Google
Google rolls out notebooks in Gemini as synced “personal knowledge bases” with NotebookLM, letting users organize chats and files for projects, initially for paid subscribers on the web.
Singularity Soup Take: The chatbot war is becoming a workflow war, whoever owns your projects, files, and context owns your “default brain,” and that’s the real product.
Enterprise and safety blueprints (because of course)
The next phase of enterprise AI — OpenAI
OpenAI argues enterprise adoption is accelerating, sketches a strategy around agents plus a unified workplace interface, and offers the kind of numbers-and-momentum narrative that CFOs can frame.
Introducing the Child Safety Blueprint — OpenAI
OpenAI publishes a policy blueprint focused on AI-enabled child safety harms, proposing legal modernization, better reporting coordination, and safety-by-design measures across the ecosystem.
Relevant Resources
Your AI Privacy Guide — Handy background for the “log in, share context, trust us” era.
Google Gemini — What Gemini is and where it fits if notebooks become the new default workflow.
Today's Pulse: 9 stories tracked across 9 sources — OpenAI, Google Cloud, CNBC, 9to5Mac, TechCrunch, BBC News, 404 Media, Transparency Coalition, Google Blog