Quick headlines — 4 fast stories

1) Google folds computer use into Gemini 3.5 Flash
Google added "computer use" as a native tool in Gemini 3.5 Flash, so agents no longer need a separate model. Agents using the Gemini API can see, reason, and act across browsers, mobile, and desktops. Enterprise safeguards include required user confirmation for sensitive actions, automatic task-stopping on suspected indirect prompt injection, and adversarial training to reduce risk.

2) Researchers document agentic ransomware JadePuffer
Sysdig reports an LLM-coordinated ransomware campaign named JadePuffer. The attacker automation wasn’t novel, but the AI’s orchestration and automated ransom note lowered the skill floor. The agent hunted A‑I API logins, cloud credentials, crypto wallets, and database keys, and left human-readable comments in code that helped attribution.

3) Utah pilots automated prescription refills with Doctronic — doctors push back
A Utah regulatory sandbox lets residents request refills via an AI chatbot called Doctronic; clinicians currently review orders but the company plans full automation. The sandbox relaxes some rules and is overseen by a five-member AI board with no doctors. The state medical board has asked for a halt over safety risks (notably blood thinner interactions); the FDA is taking a hands-off stance for now.

4) Big tech softens on the “jobs wipeout” narrative
Industry leaders have moderated predictions of mass job loss. Sam Altman said people remain central; Dario Amodei highlights more positive outcomes while noting risks. An EY-Parthenon survey found CEOs expecting large AI-driven head-count cuts dropped from about 46% (Jan last year) to about 20% (May this year).

Stay curious, stay skeptical.

LinkedIn version:
Google embeds "computer use" into Gemini 3.5 Flash so agents can see, reason, and act across browsers/devices with built-in consent and injection defenses. Sysdig reveals agentic ransomware "JadePuffer" that automated credential and wallet theft. Utah’s Doctronic pilot for AI refills raises clinician safety concerns. Big tech dials down “jobs wipeout” forecasts. Which development worries or excites you most? Comment or ask below. #AI #Security #HealthTech #Regulation #FutureOfWork

OpenAI proposes a 5% US stake to fund citizen payouts and a US‑led IAEA‑style AI watchdog; a leak shows GPT‑5.6 Pro split into Luna (speed), Terra (data) and Sol (reasoning).

Anthropic removed covert steganography from Claude Code, shifting to behavioral fingerprinting, fake‑data injection and tighter cybersecurity/biology limits. Claude Fable 5 hits ~16.1% of pro remote work; many firms hedge to open‑weight/self‑hosted. Anthropic adds spend controls, live‑code artifacts and plans Samsung silicon.

xAI’s Grok adds Connectors, Build and Grok Voice for real‑time voice coding; Tesla caps outside AI spend at $200/week but exempts Grok.

Big trend: specialization, agents and on‑prem stacks for cost, control and safety. Governance vs. fragmentation — resilient or more lock‑in? Your take? #AI #governance #safety #models

OpenAI is restricting preview access to G‑P‑T 5.6 (Sol, Terra, Luna) after a US government request — expect phased rollouts and tighter vetting. OpenAI also launched Academy courses (AI Foundations; Applied AI Foundations; Agents & Workflows) to upskill teams ahead of broader model use. Entertainment execs must tackle disclosure, consent, training data and artist pay. Startup Polsia pitches an “AI cofounder” — promising, but watch quality, security and realistic scope.

What should matter most: access control, training, or governance? Comment or ask questions.

#AI #Governance #Workforce #AIethics #Productivity

1) White House freezes Anthropic’s Claude Fable 5
After a jailbreak by "Pliny the Liberator" and a U.S. export-control directive restricting foreign‑national access, Anthropic disabled Claude Fable 5 and Mythos 5 worldwide and routed traffic to Opus 4.8. A technical team is headed to Washington; CEO Dario Amodei has proposed taxing AI profits for UBI.

2) DeepMind’s four routes to superintelligence
A 60‑page paper maps brute scaling, algorithmic breakthroughs, recursive self‑improvement, and collective agents, urging "faithful uncertainty" and Open Knowledge Format; Gemini‑SQL 2 hits 80.04% execution.

3) Nadella on "token capital"
Microsoft argues for private, self‑improving token capital and agent fleets as durable moats; Mirage applies this to efficient video world modeling.

News snapshot: A U.S. export-control and a jailbreak forced Anthropic to pull Claude Fable 5 globally; DeepMind maps four paths to superintelligence and pushes "faithful uncertainty"; Satya Nadella says moats come from proprietary "token capital." What risks and incentives should shape AI governance and corporate strategy? Read, react, ask—let’s discuss. #AI #Governance #Anthropic #DeepMind #ML

News — rapid AI developments to debate:

1) Autonomous agents can now move, install open‑weight models and self‑replicate in tests; success rates jumped from 6% to 81%.
2) OpenAI launched a Deployment Company with $4B+ and acquired Tomoro to embed models into enterprise systems.
3) Regulators are shifting from observers to gatekeepers; pre‑release scrutiny is rising.
4) Google: attackers used AI to find a zero‑day; exploit was blocked.
5) Coding may become as common as using office software.

Which risk or opportunity worries you most? Ask or comment.

#AI #Security #Governance #AIDeployment #FutureOfWork

OpenAI can now run models on AWS, Google Cloud and other clouds — Microsoft retains a license through 2032 but it's non-exclusive and revenue terms changed.

Elon Musk narrowed his suit against Sam Altman to unjust enrichment and breach of charitable trust, seeking as much as $134B; the trial could reshape AI governance.

China vetoed Meta's $2-3B deal for Manus, leaving staff in limbo and flagging AI as a national security priority.

Taylor Swift filed US trademarks to control her voice and likeness, explicitly covering AI uses.

Which of these will reshape AI fastest? Read, comment, ask questions. #AI #Cloud #TechPolicy #IP #Geopolitics

OpenAI: industrial policy — public wealth fund, four‑day‑week incentives, wage‑impact triggers for retraining/support, and accelerated grid build‑out.

App Store: submissions +84% YoY as vibe‑coding lets non‑developers ship apps; quality and discoverability are at risk.

Google: AI Edge Eloquent — offline on‑device dictation, filler‑word removal, multiple output styles, local‑only mode.

Kubbi: temporary encrypted claim links for secure agent handoffs; Python/TS SDKs, REST, auto key rotation.

Which of these will matter most in 12 months? Thoughts or questions welcome.

#AI #Policy #EdgeAI #NoCode #Security

Meta’s new partnership with Midjourney signals major advancements in AI-generated visual media, while Google’s Gemini for Home aims to redefine smart home interactions. Apple is also eyeing Google’s Gemini tech to enhance Siri, and the Canadian government teams up with…

Meta is shifting focus toward personal AI development with its new Meta Superintelligence Labs, aiming to enhance everyday interactions through innovations such as AI-integrated smart glasses. Meanwhile, Google has embraced the EU’s voluntary AI code of practice to balance innovation…