Won't Get Fooled Again
The AI industry keeps describing itself in world-historical terms. Some of that is earned. The money, compute, policy attention, labor-market anxiety, and security implications are real.
But today’s news points to a more grounded version of the story: AI will change everything and nothing at once. It will reshape institutions, workflows, fraud, infrastructure, customer service, cyber defense, and labor policy. And for most people, it will arrive as another Tuesday with a new boss, same as the old boss.
• Amazon reportedly played the court-intrigue role in the Fable/Mythos shutdown. Axios reported that Amazon CEO Andy Jassy called Treasury Secretary Scott Bessent to raise concerns that Anthropic’s Mythos and Fable models could be jailbroken; Axios separately reported that an urgent Amazon report helped trigger a White House scramble that ended with Anthropic taking the models offline. TechCrunch, citing Wall Street Journal reporting, also reported that Jassy told Bessent and other officials that Amazon researchers had used Claude Fable 5 to obtain information that could be useful in cyberattacks. Maybe that is responsible disclosure. Maybe it is also what old power has always looked like: a major investor, cloud partner, platform gatekeeper, and strategic rival whispering into the king’s ear, followed by a blunt sovereign response. AI may change the tools of power. It has not changed the choreography. (axios.com)
• The “AI changes everything” camp now has its public-benefit doctrine. OpenAI’s recent “Built to benefit everyone” plan uses electricity as its analogy: a technology that changed ordinary life unevenly, practically, and over time. OpenAI’s own newsroom also shows the company moving on several fronts at once: confidential S-1 submission, Oracle cloud access, economic research, Codex, memory, and product updates. That is not apocalypse. It is infrastructure strategy. (openai.com)
• Anthropic’s economic-policy paper is the other side of the same Tuesday. Anthropic’s June 2026 labor-policy framework says some displacement may be an intrinsic consequence of AI, but emphasizes preparation: tracking AI’s effects, modernizing unemployment insurance, building workforce training into deployment, redesigning early-career roles, and considering broader support if disruption deepens. The important signal is not that Anthropic has solved the labor question. It is that AI firms are now writing policy memos for the social systems their products may strain. (www-cdn.anthropic.com)
• Salesforce buying Fin is what AI transformation actually looks like in enterprise software: customer support gets folded into the incumbent stack. Salesforce agreed to acquire Fin, formerly Intercom, for about $3.6 billion, with the stated goal of strengthening Agentforce’s service-agent capabilities. That is not a robot revolution. It is the old enterprise pattern: buy the workflow, integrate the data, sell the platform, automate the department. For the customer, it may feel like faster support. For the worker, it may feel like fewer seats. For the buyer, it becomes another line item in the CRM budget. (salesforce.com)
• The OpenAI–Nvidia compute deal shows that even “intelligence” still runs on steel, power, chips, and capital. Nvidia said its strategic partnership with OpenAI contemplates deploying at least 10 gigawatts of Nvidia systems for OpenAI’s next-generation AI infrastructure, with Nvidia intending to invest up to $100 billion as systems are deployed. The first gigawatt is targeted for the second half of 2026. The mystical language around AI often obscures the practical reality: this is an energy, real estate, supply-chain, semiconductor, and financing story. (openai.com)
Orthogonal Take
The day’s signal is not that AI changes everything or that AI changes nothing. It is both.
At the top of the stack, AI is already a fight over public markets, export controls, cyber capabilities, infrastructure, labor policy, and enterprise software consolidation. At the bottom of the stack, it shows up as a support bot, a new setting, a faster workflow, a suspicious text message, a better scam filter, a resume helper, or a customer-service agent that never says it is tired.
That is the more historically grounded frame. AI will not repeal the human condition. It will not end ambition, fraud, bureaucracy, labor conflict, institutional self-preservation, national-security competition, or the old struggle over who controls the tools. It will make those things faster, cheaper, more scalable, and harder to see.
For most people, the future will not arrive as a trumpet blast. It will arrive as a product update, a policy change, a denied ticket, an automated review, a better diagnosis, a worse scam, a cheaper service rep, a new subscription tier, or a meeting where someone says the team can “do more with less.”
Another Tuesday.
Meet the new boss, same as the old boss.