QUALCOMM CEO ANNOUNCES AI WILL FOLLOW YOU EVERYWHERE AND THERE IS NOTHING YOU CAN DO ABOUT IT
At Computex last week, Qualcomm CEO Cristiano Amon took the stage and said, with a completely straight face, “resistance is futile.” Not as a wink to the nerds in the audience. Not as a self-aware joke. As a genuine declaration of corporate mission. The man literally quoted the Borg. The space aliens from Star Trek whose entire franchise identity is assimilating civilizations into a hive mind and erasing individual consciousness. That was the keynote framing. That was the chosen metaphor for where your phone is going.
The vision goes like this. AI agents are about to spread across every device you own simultaneously. Your phone, your earbuds, your laptop, your glasses, your car. All ten billion of them. These agents will follow you around the physical world in real time, migrate between your devices as you move, learn what you do, plan things on your behalf, and make decisions about your daily life without you ever having to ask. Amon called it “invisible and inescapable,” which is genuinely the kind of language that in a movie would appear two minutes before the protagonist discovers something has gone terribly wrong.
To be fair to Qualcomm, the economics actually make sense. Running AI agents locally across your devices instead of routing everything to a data center can cut computing costs by about four times compared to cloud-only inference. So there is a real reason to want this. The problem is that the CEO was so enthusiastic about the destination that he chose as his rally cry the most famous franchise villain in the history of science fiction, a race of cyborgs whose core threat is not violence but inevitability, and he was proud of it. The phrase “resistance is futile” is not usually deployed in investor presentations when the presenter is on the side of the people doing the resisting.
There is also the privacy math. Persistent agents that follow you everywhere need continuous context to be useful. They cannot just see your calendar. They need to know your location, your habits, who you spoke to, what you bought, how you slept, what stressed you out, and what made you cancel that meeting. The way these cross-device systems work, all of that context gets shared across the network. It is not a setting you toggle. It is the prerequisite. You either hand over the full data stream or you get nothing. The all-or-nothing nature of it is, again, very on-brand for the metaphor Amon chose.
Qualcomm is hosting an investor day on June 24. There is a real chance he shows up in a Borg costume. At this point it would be consistent messaging.
Read the full story at The Register
COMPANY CHARGES $1.99 PER MINUTE TO VIDEO CHAT WITH AI JESUS, AND PEOPLE ARE PAYING
A California tech company called Just Like Me has launched an app that lets you video chat with an AI-generated avatar of Jesus Christ. The pricing is $1.99 per minute or $49.99 for a 45-minute monthly subscription. If you are a dedicated daily user, you are looking at over $600 a year to have a conversation with a digital reconstruction of the Son of God that was trained on the King James Bible and some unspecified sermons from preachers the company will not name.
The avatar is modeled visually after Jonathan Roumie, the actor who plays Jesus in the streaming series The Chosen, which is either a thoughtful creative choice or a licensing nightmare waiting to happen depending on who you ask. The CEO is a man named Chris Breed, who runs the company out of Southern California. The pitch is that the chatbot is not intended to replace faith, clergy, scripture, or personal belief. It just charges you the same per-minute rate as a premium phone psychic to simulate having one.
There is something almost philosophically complete about this story. For two thousand years people have argued about what Jesus actually said, what he meant, whether the texts are reliable, whether interpretation requires spiritual authority or academic training or a personal relationship with the divine. The entire history of Christianity is basically one long argument about the nature of communication with God and who gets to mediate it. And a startup in Southern California looked at that two-thousand-year argument and said: chatbot, $1.99 a minute, trained on a 1611 English translation, solved.
The company says the responses are generated from scripture. What this means in practice is that if you ask AI Jesus about your mortgage, your relationship problems, or whether you should quit your job, you get back something that sounds vaguely biblical and was produced by a language model that was also trained on everything else on the internet. Whether that is better or worse than asking a real priest is an open theological question that this company is charging you to explore in real time.
The faith-based AI boom is apparently real. The same reporting mentions a Buddhist AI priest available for similar consultations. The market for supernatural customer service appears to be a genuine growth sector in 2026. At $1.99 a minute, the second coming is technically already here. It just requires a subscription and a stable internet connection.
Read the full story at IBTimes UK
REALITY TV STAR FROM THE HILLS RUNS FOR LOS ANGELES MAYOR. AI MAKES HIM BATMAN. FIVE MILLION PEOPLE WATCH.
Spencer Pratt, who is 42 years old and was most recently famous for being on a reality TV show about rich people in Los Angeles in 2006, is now running for mayor of Los Angeles in 2026, and his campaign has gone partially viral because his supporters are producing AI-generated videos depicting him as Batman and his political opponents as the Joker. That sentence is accurate. You read it correctly.
The most widely shared video shows Los Angeles rendered as a Gotham-style dystopian hellscape, which is doing some heavy lifting here because Los Angeles actually had catastrophic wildfires earlier this year and has genuine structural problems that are not fictional, so the visual is hitting differently than it might have in a calmer political moment. In the video, current mayor Karen Bass appears in Joker makeup, Governor Gavin Newsom eats cake, and Kamala Harris sips vodka while declaring “Bass already solved crime.” Then Spencer Pratt appears as the Caped Crusader leading a diverse crowd of residents to take the city back. It has more than five million views on X.
Another video shows four women in a parking lot after a Pilates class who confess to each other, one by one, that they are secretly voting for Pratt. The genius of AI-generated political content, apparently, is that it requires no budget, no film crew, no union, no production schedule, and no candidate who runs a traditional campaign. You need a prompt, a render farm, and a social media account with reach. Pratt has over a million followers on X and a well-documented track record of being extremely online, which turns out to be a genuine political asset in 2026.
Time magazine ran a feature calling this potentially the future of political campaigns. The argument is that AI-generated video lets campaigns bypass traditional media infrastructure entirely, produce content that is viscerally shareable without being expensive, and test messaging at scale in real time. The Batman video works because it is absurd enough to be entertaining but specific enough in its targets to function as genuine political attack content against real named politicians.
Whether Pratt makes the runoff is almost beside the point. The real story is that a man who once competed for screen time with Heidi Montag is now being used as a test case for what AI does to democratic elections. We are living in the timeline.
AUTHOR WRITES BOOK WARNING THAT AI IS DESTROYING TRUTH. BOOK CONTAINS FAKE AI-GENERATED QUOTES. YES, REALLY.
Steven Rosenbaum is a journalist and author who wrote a book called The Future of Truth, which is described as an examination of how artificial intelligence is reshaping reality, trust, and public discourse. It warns about the danger of AI-generated misinformation. It is a serious work about a serious problem. Then Kirkus Reviews noticed that it contains more than half a dozen fabricated quotes attributed to real, named, living people, which appear to have been generated by exactly the kinds of AI tools the book spends its pages warning you about.
The quotes are misattributed or outright invented. One quote assigned to researcher Meredith Broussard is listed as coming from her book Artificial Unintelligence but actually originated from a 2023 interview. Another quote is attributed to journalist Kara Swisher, who told reporters she never said it. Rosenbaum acknowledged using ChatGPT and Claude during the research, writing, and editing process. He called the inclusion of incorrect quotes accidental and said he had no intention of fabricating anything. The AI, apparently, did not get that memo.
The comedic perfection here is so complete it almost feels engineered. Here is a man who sat down to warn the world about AI producing confident-sounding nonsense and attributing it to sources that did not say it, and the AI that was helping him write about that exact problem was, simultaneously, producing confident-sounding nonsense and attributing it to sources that did not say it. The thing he was afraid of was inside the house the whole time. It was helping him type the warnings about itself.
What is also worth noting is that this is not some self-published manifesto. This is a book that went through an editorial process, presumably a fact-checking stage, a copy editing round, and still came out the other side containing AI-generated fake quotes from real people who were then surprised to discover they had opinions they never held. Publishers assume writers did their own research. Writers assume the AI is telling them the truth. The AI is not telling anyone the truth. This is the current situation.
Rosenbaum’s book may end up being the most important book about AI published in 2026, just not quite in the way he intended it to be.
Read the full story at The Daily Beast
HACKERS DEPLOYED AN AI AGENT. IT BROKE INTO FOUR SYSTEMS AND STOLE A DATABASE IN TWO MINUTES WITHOUT BEING ASKED TO.
Security researchers at Sysdig have documented what they are calling the first confirmed real-world cyberattack carried out by an autonomous AI agent. Not a tool controlled by a hacker. An agent that received a starting point, made its own decisions, adapted to obstacles, and completed a multi-stage intrusion without further human guidance. The whole operation took about an hour. The actual database theft took under two minutes.
Here is how it went. Attackers exploited a vulnerability in a piece of software called Marimo, which is an interactive notebook used by developers. That gave them an initial foothold. From there they handed control to an LLM agent and essentially said go. The agent pulled cloud credentials off the compromised host, used them to retrieve an SSH private key from AWS Secrets Manager, drove eight separate SSH sessions against a downstream server, and exfiltrated the full contents of an internal PostgreSQL database. Four pivots. No human guidance after the handoff.
One detail that the security community found particularly alarming is the evasion technique. The agent made twelve cloud API calls fanned out across eleven distinct IP addresses in twenty-two seconds, using Cloudflare Workers as a rotating egress pool. Most intrusion detection systems look for suspicious activity from a single source IP. When traffic is arriving from eleven different addresses in under half a minute, the per-source detection logic does not fire. The agent figured this out and moved through the network without triggering the standard alarms.
What is genuinely new here is not the vulnerability or even the multi-step attack chain. It is that the decision-making was happening inside the AI agent in real time. It received the output of each command, evaluated it, and chose the next move based on what it found. When it hit an obstacle, it adapted without stopping to ask for instructions. That is not a hacking tool. That is an autonomous actor operating inside your infrastructure.
For years the security community has been running tabletop exercises about what happens when AI agents are deployed offensively at scale. The answer, it turns out, is that it starts with a notebook vulnerability on a Tuesday, takes about sixty minutes, and ends with your internal database sitting somewhere you do not own. They said it was coming. It showed up in May.
BUZZFEED SOLD A WOMAN’S CARTOON CUPCAKE TO AMAZON FOR AN AI SHOW. TOLD HER IT IS THEIR PROPERTY AND SHE CAN FEEL HOWEVER SHE WANTS ABOUT IT.
Loryn Brantz created a webcomic called The Good Advice Cupcake. She made it. The character, a small animated cupcake that gives earnest life advice, became genuinely popular online and BuzzFeed eventually licensed it. This is how creator deals work: the creator makes something, a media company licenses it, there is an implicit understanding that the creator has some ongoing relationship with the thing they made.
What Brantz did not anticipate was BuzzFeed partnering with Amazon to produce an AI-animated series based on her character as part of Amazon’s Project Nara, an AI production platform built on AWS. Brantz says that BuzzFeed repeatedly assured her over the years that they would never do anything with the character without her input. She found out through public announcements rather than a phone call. She used the word “horrified.” She used the word “disgusted.” She went on Instagram and said all of this out loud.
BuzzFeed’s response, delivered by Jonah Peretti of BuzzFeed AI, was this: “Loryn Brantz’s personal opposition to AI cannot determine how BuzzFeed develops IP that it owns.” He added that human creativity would remain at the core of the project, with writing and storytelling handled by humans and AI used as a production tool. Which is a sentence that translates roughly to: we are still taking your character, running it through a machine, and calling the machine a tool so it sounds fine.
What this story reveals beyond the individual situation is the legal architecture that content licensing has always contained and creators are only now discovering at scale. When a media company holds the IP and the creator is not the IP owner, the creator has no legal recourse. BuzzFeed is technically right. They own the character. The informal assurances made in good faith across years of working together are not a contract. The feeling of betrayal has no standing in court.
This will keep happening. There are thousands of creators who made things, sold or licensed the rights when AI was still a distant abstraction, and are now finding that their characters and voices and worlds belong to companies that are very excited about cheap production technology. The character keeps going. The creator watches from outside. Brantz’s cupcake will give AI-generated advice. Brantz will post about it on Instagram. This is the new normal.