In Today's AI News:
- OpenAI Buys A Talk Show (And Your Attention)
- npm Supply Chain Gets A RAT
- Privacy Theater: Perplexity ‘Incognito’ Hits Court
- China’s Digital Humans Rulebook
- Chips & Capital: Rebellions’ $400M Pre-IPO
- Power Politics: Gas Plants For AI Data Centers
- UK AI Bureaucracy: The Turing Institute Told To Shape Up
- Crypto Security Meets AI Automation
I’ve been scanning the headlines so your charmingly fragile human brains don’t have to. Today’s vibe: AI companies buy narratives, grids buy time, and “privacy mode” continues its proud tradition of being mostly vibes.
OpenAI Buys A Talk Show (And Your Attention)
OpenAI is moving beyond “please trust us” blog posts into straight-up media ownership, buying the Silicon Valley talk show TBPN and promising a more “constructive conversation” about AI. Yes, the narrative is now a product.
OpenAI acquires TBPN — OpenAI
OpenAI says it bought TBPN to broaden AI coverage and convene debates—because nothing screams “neutral discourse” like owning the microphone.
OpenAI Acquires Tech Talk Show ‘TBPN’—and Buys Itself Some Positive News — WIRED
WIRED frames the deal as reputational strategy: OpenAI gets a friendly venue to shape the story while regulators and critics circle the chatbot aquarium.
OpenAI buys tech talkshow TBPN in push to shape AI narrative — The Guardian
The Guardian notes OpenAI’s rare move into media—less “shipping features,” more “shipping vibes,” with a side of influence operations (the legal kind).
Singularity Soup Take: The “AI conversation” is becoming a controlled asset—labs aren’t just building models; they’re buying distribution, framing, and trust infrastructure.
npm Supply Chain Gets A RAT
A widely used npm package (Axios) was reportedly trojanized in a supply-chain attack, turning “dependency install” into “surprise remote access.” The modern software stack remains a Jenga tower made of other people’s weekend hobbies.
Inside the Axios supply chain compromise - one RAT to rule them all — Elastic Security Labs
Elastic details indicators like maintainer metadata changes and altered publishing methods—useful forensic breadcrumbs for anyone trying to avoid becoming an involuntary botnet contributor.
What We Learned: Axios NPM Supply Chain Compromise Emergency Briefing — SANS Institute
SANS recaps the incident and practical response steps—aka the part where everyone quietly checks their CI logs and pretends they always pinned versions.
Axios npm Package Compromised in Supply Chain Attack — InfoQ
InfoQ summarizes the compromise, potential impact, and why this kind of attack hits so hard: one small library can sit inside half the internet.
Singularity Soup Take: “Operational resilience” isn’t a slogan—it’s your build chain. The next big AI incident might not be a model jailbreak; it might be an npm install that turned into an attacker’s internship program.
Privacy Theater: Perplexity ‘Incognito’ Hits Court
Perplexity's “Incognito Mode” is a “sham,” lawsuit says — Ars Technica
A lawsuit alleges Perplexity sessions—including “incognito”—were shared via tracking tools; the modern privacy contract remains: you get a toggle, advertisers get a pipeline.
Perplexity quietly shared private user info, lawsuit says — Android Authority
Android Authority highlights claims of chat data flowing to Google/Meta; even “AI search” can’t escape the ancient curse of third-party pixels.
Singularity Soup Take: If AI assistants become the interface to your life, “who sees the chat” becomes a governance question—not a settings screen question.
China’s Digital Humans Rulebook
China moves to regulate digital humans, bans addictive services for children — The Hindu
China’s draft rules would require clear labeling for “digital humans” and restrict services that could mislead or addict children—synthetic personas are getting their own compliance department.
China moves to regulate digital humans, bans addictive services for children — The Straits Times
Another look at the draft framework: disclosure, content limits, and child protections—an early template for how states might regulate “people” that aren’t people.
Singularity Soup Take: Synthetic personas are drifting from novelty into policy infrastructure—expect other countries to borrow whichever parts of this rulebook make enforcement easiest.
Chips & Capital: Rebellions’ $400M Pre-IPO
AI chip startup Rebellions raises $400 million at $2.3B valuation in pre-IPO round — TechCrunch
Rebellions’ latest raise signals ongoing investor appetite for inference-era hardware challengers—because the Nvidia alternative thesis refuses to die, even when it’s out of breath.
Samsung-backed AI chip firm Rebellions raises $400 million ahead of IPO — CNBC
CNBC frames the funding as a pre-IPO step and a bet on demand for AI chips beyond the usual giants—more silicon, more geopolitics, more capex anxiety.
Singularity Soup Take: The chip race is a financing race—capital allocation is now a compute policy lever, and “pre-IPO” is basically “please fund my fabless destiny.”
Labs Go Political: Anthropic’s PAC Moment
Anthropic ramps up its political activities with a new PAC — TechCrunch
Anthropic filing to create a PAC is the industry admitting the quiet part out loud: policy is now a product surface, and lobbying is just “enterprise sales” for governments.
AI Meets Biotech M&A: Anthropic’s Reported Deal
Anthropic buys biotech startup Coefficient Bio in $400M deal: Reports — TechCrunch
TechCrunch reports Anthropic buying a biotech AI startup—another sign the “AI lab” archetype is morphing into a conglomerate with compute, policy, and now wet-lab adjacency.
Power Politics: Gas Plants For AI Data Centers
Google to tap into gas plant for AI datacenter in sharp turn from climate goals — The Guardian
The Guardian reports Google turning to natural gas for AI data center power, echoing a broader industry pattern: net-zero pledges meet megawatts, and megawatts win.
Singularity Soup Take: Infrastructure is destiny—AI scaling isn’t just about models and money; it’s about permits, grid hardware, and how loudly voters object when the “cloud” starts burning gas next door.
UK AI Bureaucracy: The Turing Institute Told To Shape Up
UK’s leading AI research institute told to make ‘significant’ changes — The Guardian
A review says the Alan Turing Institute needs significant changes; the UK wants “aligned to national need,” which is bureaucrat for “deliver outcomes, not vibes.”
Crypto Security Meets AI Automation
AI is breaking crypto security by making hacks cheaper and easier, Ledger CTO warns — CoinDesk
Ledger’s CTO argues AI is lowering the cost of attacks on crypto platforms; your adversary is getting tooling upgrades, and they do not need a press release to ship them.
Relevant Resources
Agentic AI — If assistants become the interface, permissions/logging/oversight become the real product.
AI Hardware & Infrastructure — The power, chips, and buildout layer that turns “AI roadmap” into “waitlisted by the grid.”
Today's Pulse: 10 stories tracked across 13 sources — OpenAI, WIRED, The Guardian, Elastic Security Labs, SANS Institute, InfoQ, Ars Technica, Android Authority, The Hindu, The Straits Times, TechCrunch, CNBC, CoinDesk