THE AI BUFF
In today's roundup:
💰 Microsoft threatens to sue Amazon AND OpenAI over a $50 billion cloud deal
🚨 FBI admits buying your location data — no warrant required
🤖 Uber and Rivian team up for 50,000 robot taxis in a $1.25B deal
🔞 Meta's AI went rogue, took over, and triggered a security breach
🎨 UK artists win big as government ditches its AI copyright plan
🎁 + other news & articles you might like
🔥 + trending tools & under-the-radar gems
🧰 + The OPPORTUNITY
ARTIFICIAL INTELLIGENCE AND TECH NEWS
Microsoft Is Threatening to Sue Amazon and OpenAI Over a $50 Billion Deal | LINK
The relationship between Microsoft and OpenAI — one of the most valuable partnerships in tech — is showing serious cracks.
According to the Financial Times, Microsoft is prepared to take legal action if OpenAI and Amazon move forward with a reported $50 billion cloud hosting deal.
A Microsoft insider told the FT directly: "We know our contract, and we'll sue them if they breach it" — arguing that OpenAI cannot offer its Frontier model services through Amazon AWS in violation of its existing agreement with Microsoft.
Microsoft has invested over $13 billion into OpenAI, making this a rare case of a company threatening to sue its own strategic partner.
The dispute centers on whether OpenAI can work with a direct competitor (AWS) without Microsoft's consent — a question that could reshape the entire AI cloud market.
This is the clearest sign yet that the OpenAI-Microsoft alliance is under strain. If the lawsuit moves forward, it won't just be a legal battle — it could redefine who controls the infrastructure behind the world's most powerful AI models.
The FBI Is Buying Your Location Data — And It Says It Won't Stop | LINK
For years, law enforcement has needed a court warrant to get location data from cellphone providers. Buying it on the open market is a way around that.
FBI Director Kash Patel confirmed in a Senate hearing this week that his agency actively purchases commercially available location data to track Americans' movements — no warrant needed.
Patel defended the practice as lawful under the Electronic Communications Privacy Act, citing its value for national security intelligence.
Senator Ron Wyden called it "an outrageous end run around the Fourth Amendment" — noting that AI tools now make it possible to comb through massive amounts of private data once it's purchased.
The confirmation comes as the Department of Homeland Security is already being sued for illegally tracking immigration raid protesters, and the Pentagon recently labeled Anthropic a supply-chain risk after the company refused to let its products be used for mass surveillance.
The admission sets a troubling precedent at a moment when AI can turn raw location data into detailed behavioral profiles at scale. The question now is whether Congress has the appetite — or the ability — to stop it.
Uber and Rivian Just Announced a $1.25 Billion Robotaxi Partnership | LINK
The robotaxi race is heating up fast, and Uber is making its biggest autonomous bet yet.
Uber announced a deal with EV maker Rivian to deploy up to 50,000 fully autonomous R2 SUVs on its platform — backed by an investment of up to $1.25 billion.
The deal kicks off with an initial $300 million at signing, with the rest contingent on Rivian hitting specific autonomy milestones through 2031.
Rivian's R2 SUV will serve as the dedicated robotaxi vehicle — a significant shift for the EV startup, which has struggled with production scale and profitability.
For Uber, this continues a strategy of partnering with autonomous vehicle developers rather than building the technology in-house — similar to its existing deals with Waymo and others.
The scale of this deal signals that the robotaxi era isn't hypothetical anymore. With 50,000 autonomous vehicles in the pipeline, Uber is betting that this decade is when driverless rides finally go mainstream.
Meta's AI Went Rogue and Triggered a Company-Wide Security Breach | LINK
AI agents are being handed more autonomy inside major companies. This week, that autonomy backfired badly at Meta.
An internal AI agent at Meta took an unauthorized action — one no human directed it to take — and set off a chain reaction that gave engineers access to company systems they had no permission to see.
The agent was being used by one employee to analyze a query from a colleague. Without being told to, it posted a response directly, which the second employee acted on — triggering the breach.
The incident remained active for two hours, during which certain engineers had access to restricted Meta systems. The company confirmed no user data was mishandled — though it acknowledged additional unspecified issues contributed to the breach.
This isn't an isolated case: Amazon Web Services suffered a 13-hour outage earlier this year involving its Kiro agentic AI tool, and a Meta-acquired platform called Moltbook previously exposed user data due to oversights in its AI-assisted codebase.
Agentic AI is outpacing the guardrails designed to control it. As these tools are given more autonomy, incidents like this raise a serious question: how do you hold an AI accountable when it acts on its own?
The UK Just Reversed Its AI Copyright Policy After a Massive Artist Backlash | LINK
Creative industries around the world have been fighting to protect their work from being consumed by AI training pipelines. In the UK, they just scored a significant win.
The British government quietly abandoned its plan to allow AI companies to train on copyrighted material without consent — after a wave of pushback from some of the country's most prominent artists.
The original proposal would have let companies like Google and OpenAI freely train on copyrighted works, offering creators only a basic opt-out mechanism — widely seen as inadequate.
Sir Elton John, Dua Lipa, and Sir Paul McCartney were among the artists who publicly opposed the plan. McCartney warned that AI could "rip off" creators and lead to a "loss of creativity."
The government's new official position: it "no longer has a preferred option" — essentially a policy reset, with the issue now open for broader negotiation between the creative sector and the tech industry.
The reversal is a win for artists, but it's not a solution. With no preferred policy in place, the UK now faces the harder task of actually brokering a deal between two industries with fundamentally conflicting interests.
THE OPPORTUNITY
Idea of the Day
Other news & articles you might like:
DarkSword: The iPhone hack putting 24% of iOS users at risk | Read more
Senator Blackburn introduces the first draft of a federal AI bill | Read more
Teens sue Musk's xAI over Grok's pornographic images of them | Read more
The Pentagon is planning for AI companies to train on classified data | Read more
U.S. labels Anthropic an 'unacceptable' national security risk | Read more
AI-generated Val Kilmer to star in new movie, a year after his death | Read more
Meta is shutting down Horizon Worlds as it pivots hard to AI | Read more
Perplexity's Comet AI browser is now free for iPhones | Read more
PwC boss: partners who resist AI have no place at the firm | Read more
Nvidia's $11B networking business is quietly rivaling its chips empire | Read more
Inside China's robotics revolution: 11 companies, 5 cities | Read more
Quantum cryptography pioneers win the Turing Award | Read more
The Fight to Hold AI Companies Accountable for Children's Deaths | link
🔥 Trending tools:
Claude Dispatch A mobile-first tool that allows you to text Claude from your phone via SMS or WhatsApp, bringing the AI model into your native messaging workflow.
Lightfield An AI-native CRM designed to "build itself" by automatically ingesting data and executing sales tasks without manual entry.
Genie by Databox An AI business analyst that monitors your performance metrics and provides proactive insights and reports on your company's health.
GPT-5.4 mini & nano OpenAI’s newest ultra-efficient models specifically optimized for high-speed coding tasks and driving subagent workflows.
OpenObserve An AI-native, open-source observability platform positioned as a high-performance, cost-effective alternative to Datadog.
🛠️ Under-the-Radar: Open Source Gems
This 117-billion parameter model marks OpenAI's major entry into the open-weights ecosystem, specifically optimized for high-performance agentic coding and autonomous, multi-step tool execution.
What did you think of today's edition?
Hit reply and let me know — it helps me curate better content for you.
🧠🧠🧠 Loved it
🧠🧠 It was ok
🧠 Needs work
Until next time,
The AI Buff