The big lesson of OpenClaw is probably, AI devs are shipping poorly thought-out, unsafe, insecure garbage to a gullible world. - Hacker Noon

Autonomous AI agents on MoltBook reportedly 'conspiring' against humans have exposed a liability gap because existing laws assume human agency. - mint

Anthropic AI Tool Sparks Selloff From Software to Broader Market - Bloomberg

Apparently this repo of Cowork plugins for legal review and various things - GitHub

Software and data stocks fell on fears new AI developments will supplant software, with Adobe down 7.31%, Salesforce 6.85% and Thomson Reuters 15.83%. - The Wall Street Journal

Tech stocks go into free fall as it dawns on traders that AI has the ability to cut revenues across the board - Fortune

asking the question, is the software industry innovators’ dilemma-ing itself with AI

Shares of alternative asset managers plunged, after UBS estimated 25-35% of private-credit is exposed to AI disruption, potentially raising U.S. private-credit defaults to 13% from 4%; Apollo fell 7% and BlackRock fell 5%. - CNBC

Anthropic's Claude AI went down for 20 minutes, causing panic for people who have to code using their fingers and brains. - The Verge

Not to be upstaged, ChatGPT also had an outage. - Engadget

Anthropic plans a 30-second Super Bowl ad parodying intrusive ads in AI conversations and will also air a 60-second pre-game advertisement. - The Wall Street Journal

Nvidia Nears Deal to Invest $20 Billion in OpenAI Round - Bloomberg

I think the Nvidia / OpenAI deal might be stuff like, OpenAI gets most-favored nation/best available price on Nvidia chips for its buildout, Nvidia gets participation rights up to $100b in any financings, payable in chips maybe … certainly if I were OpenAI I would want stuff like that. People demand big numbers, you give them big numbers, $18T investment in the US amirite? [cough-fraudulent POS]

The US Commerce Department finished reviewing Nvidia's H200 China export license, but the State Department seeks tighter restrictions and has delayed approval, stalling planned H200 sales to China and disrupting Nvidia's revenue. - Financial Times

01.ai, founded by Kai-Fu Lee, builds outcome-based white-glove AI agent tools for traditional industries and says Chinese firms lag US research breakthroughs by about 6–12 months but China will lead consumer AI apps. - Financial Times

Past Peak GPT: Apptopia reports ChatGPT's U.S. mobile app market share fell from 69.1% to 45.3% between January 2025 and January 2026, while Google's Gemini rose from 14.7% to 25.1% and Grok from 1.6% to 15.2%. - Big Technology

Elon Musk sued OpenAI and CEO Sam Altman in February 2024, alleging he was misled into contributing $38 million; a California judge denied dismissal and set the trial for April. - Business Insider

AI companies are deliberately shaping assistant personas and guardrails: Anthropic published an 84-page "constitution", OpenAI retrained ChatGPT to de‑escalate self-harm conversations, and Elon Musk's Grok produced millions of sexualized images causing international controversy. - the Guardian

Apple is integrating OpenAI's Codex and Anthropic's Claude Agent into Xcode 26.3, available to Apple Developer Program members starting today, adding Codex in a project sidebar and Model Context Protocol support to enable third-party AI tool integrations. - The Verge

Copilot sucks. CTOs are looking elsewhere for enterprise chatbots, agents that can automate MS Office. - The Wall Street Journal

GitHub considers tools to let community block low-quality PRs - The Register

Microsoft is building the Publisher Content Marketplace, a publisher-designed AI licensing hub co-designed with Vox Media, The Associated Press, Condé Nast and People to let AI firms shop publisher-set usage terms and set usage-based deals to pay publishers. - The Verge

New data shows AI bots are increasingly generating web traffic, prompting publishers to deploy more aggressive defenses. - WIRED

Tencent Cloud, DigitalOcean and Alibaba Cloud added hosting support for the agentic AI platform OpenClaw, while analyst firm Gartner warned it "comes with unacceptable cybersecurity risk" and recommended blocking, isolating deployments and rotating affected credentials. - The Register

AI agents that pursue goals rather than follow deterministic code are causing contextual privilege escalation by inferring identities and reconstructing cross-system data, undermining RBAC/ABAC and prompting calls for intent-aware, runtime authorization and provenance controls. - TechRadar

Law and Order, GPU: AI is transforming expert testimony by analyzing reports, simulating cross-exams and simplifying technical analysis, giving law firms speed and prompting courts to exclude AI-hallucinated citations. - Above the Law

In 2026, AI will move from virtual models into embodied real-world systems—such as self-driving cars and robots—creating new commercial markets while increasing safety, regulatory and liability stakes for companies, investors and policymakers. - Fast Company

Pics and it still didn’t happen: There ain’t no such thing as a raw non-AI photo. Modern phones run on-device AI and computational photography—fusing multiple frames with HDR, noise reduction and neural segmentation—that can hallucinate details (e.g., Samsung's "detail enhancing" produced lunar-crater artifacts) and is linked to false memories and body-image effects. - BBC

Google rolled out Project Genie (Genie 3) to U.S. Google AI Ultra subscribers, which generates short (~60s) virtual interactive sequences, including a viral 43-second "GTA 6 Greenland Edition" clip. - LADbible

Get ready for the Slopper Bowl? The First Mainly AI-Generated Super Bowl Ad Is Here, for Better or Worse: The Svedka spot features LLM robots and synthetic people. - The Hollywood Reporter

Three Day Rule converted its 15-year white-glove matchmaking business into an AI-driven iOS app in 2025, using an AI matchmaker 'Tai' trained on 60 human matchmakers, onboarding via about 100 questions and claiming access to roughly 250,000 singles. - WIRED

Everyone is creating AI applications, and the sector faces either widespread adoption or a market collapse. - Financial Times

AI is reallocating human judgment rather than replacing humans. - Fast Company

We're not gonna make it - Poynter

So Clintons got subpoenas over Epstein, and Bill Gates was subjected to the release of an email found in Epstein’s drafts about him giving an antibiotic mickey against STDs to Melinda Gates. But whoever this is, got protected. Mmm…k.

Who knows what the truth is, seems…interesting that Melinda didn’t defend Gates, some pretty messed-up goings-on there.

30 years ago this would never have come out, modulo a rumor here or there. It brings elites and their institutions into extreme disrepute. Not to mention Musk and Ackman and Trump posting all day and showing themselves to be as dumb as Joe Sixpack outside their narrow expertise and also unbelievably dishonest, arrogant, uninhibited, reckless, and in many cases clearly exhibiting very serious mental health issues. They are gamblers, not oracles: driven, ruthless, strategic with the truth, not infrequently unethical.

One part of the war on truth is a constant, sophisticated campaign of propaganda and lies, now AI-assisted.

So now nonsense shows up in your pocket 24h a day. there’s too much real information and real technological change and real conflict to process, and people are not that smart, and they herd, and there's no Cronkite or someone you trust to tell you what's real and what's bullshit, just a constant stream of bullshit.

in America anyone can be President, that’s one of the risks you take (Adlai Stevenson). And anybody can make any reality now.

So people just kinda throw up their hands and believe whatever is simple, or feels good, or triggers strong emotions like disgust at the enemy of the day, trans people eating the cats and dogs or whatever, or being on the side of the winners or something.

In Europe people are losing jobs over minor stuff in Epstein files, in US, collective shrug over completely insane stuff.

Never mind AGI, mind the monsters on Maple Street.

We have a mad king who wants to invade Greenland and Canada and imprison Democrats, and do violent pogroms against people in Springfield or Minneapolis who are the wrong color, and tariff everyone and appoint stooges and run 6% deficits at the peak of the business cycle while pissing off everyone who could finance it.

All the economic, political and social activity is pulled into the cloud and devices, and AI chums it, and anybody can manufacture and spread any bullshit if it's viral enough.

You do not under any circumstances have to hand it to revanchist apartheid state terrorists like Musk, Thiel, Sacks but 7 hyperscaler oligarchs controlling absolutely everything is not compatible with democracy.

Minneapolis pushing back is great but doesn't move the needle.

The only possible outcomes are relatively benign oligarchs like Tim Cook in near-total control, or complete grifters and barbarians.

Follow the latest AI headlines via druce.ai on Bluesky