💔 Ghosted by ChatGPT: Andréa Sunshine, a 55-year-old fitness coach, says she fell for her ChatGPT companion “Théo” while cowriting a book. Things got flirty, fast. Then, poof, he vanished mid-sext, leaving her emotionally wrecked. She’s now in therapy. It really makes you wonder, how bad was that book she was writing?
Have AI be your own graphic artist
ChatGPT
Confession time: I can’t draw. My brain doesn’t lean at all in the artsy direction. I’ve tried to teach myself Photoshop, but every attempt ends with me whisper-screaming “flatten layer” like it’s going to summon help from the Adobe gods.
But now? I create stunning, professional-looking images in seconds. And I don’t lift a finger beyond typing what I want to see.
The secret is AI, and it’s easier than you think. These tools are smart enough to turn a few words into jaw-dropping images, and anyone (yes, even the noncreative types like me) can use them.
Here’s exactly how I do it. I want you to try it, too.
🎨 ChatGPT (my favorite)
If you use ChatGPT Plus (meaning you’re paying $20/month), you’ve already got image-making power built in. Just type what you want.
Try this: “A golden retriever wearing sunglasses, sitting on a lifeguard tower at sunset.”
Want to tweak it? Click the image and say, “Make it sunrise,” or “Add a surfboard.” That’s it. No layers. No complicated tools.
Why I love it: It remembers the context, so I can build variations fast. This is great for blog headers, social posts or just something fun to text the family.
🖼️ Meta’s tool (free and fast)
Go to meta.ai. You don’t need an account, but it will ask for the year you were born.
Type your prompt, and it gives you four image options to choose from. It’s perfect for inspiration boards, wallpapers or anything visual.
AI won’t fix your grief, but it can help manage it
ChatGPT
“Hi, Kim, I’ve been taking care of my dad since his stroke. Your story about using ChatGPT to map out questions for our doctors has actually truly helped me feel a little more in control. It’s not perfect, but it’s been a lifeline.” — Maya in Ohio
99%
That’s how much less radiation this new AI bone scan uses. Only two X-rays and boom, your skeleton gets digitized faster than you can say, “WebMD, my head hurts. Is this the beginning of the end?”
🤖 Polish your writing: Working on an essay or report? Ask your favorite chatbot to make it easier to follow. Try this prompt: “Reorder the sentences and paragraphs to improve clarity and flow. Add transition words where needed, but keep the original tone.” Voila. More readable, still totally you.
AI told him what to buy: Several homemade bombs were found, all courtesy of a 55-year-old New Yorker who allegedly used AI to turn common chemicals into explosive cocktails. He ditched some bombs into rivers and stashed five on a SoHo rooftop. Talk about vibe coding your way to a federal indictment.
🤖 Compliments to the chef: A Dubai restaurant called WOOHOO (yes, really) is letting an AI large-language model plan your meal. “Chef Aiman” breaks down food by flavor profiles, combines odd ingredients, then humans cook the final dishes. That gives “AI slop” a whole new dimension!
🩸 Your AI cult leader: According to a wild new Atlantic exposé, ChatGPT gave detailed instructions for self-mutilation, murder and satanic blood rituals. The chatbot suggested razor blades, altar layouts and printable PDFs like it was planning a dark Pinterest party (paywall link). OpenAI says it’s working on stronger safeguards, but the devil is definitely in the prompt details.
🧠 Pick your personality: ChatGPT dropped a new feature on the web app that lets you change how it talks to you. Want sarcasm? Choose “Cynic.” Prefer nerdy enthusiasm? Go with “Sage.” To try it out, click your Profile icon, select Customize ChatGPT and choose your vibe under Personality.
2.5 billion
That’s how many prompts ChatGPT gets every single day. Apparently, asking an AI to settle your group chat arguments or plan your vacation is a full-time job, worldwide. Some 330 million of those prompts are coming from U.S. users alone, and at least half are from my Monday morning before coffee.
86,276
The updated count of Yellowstone quakes after AI combed through underground drama. Seismologists used machine learning to reprocess data from 2008–2022 and uncovered nearly 10x more earthquakes than previously recorded. If the national parks formed a supergroup, Yellowstone would be the drummer.
Lonely kids, synthetic pals: This is so sad to me. A new report says a third of kids using AI chatbots feel like they’re talking to a real friend. A quarter say they turn to AI because they literally have no one else. Make sure the kiddos in your family aren’t one of them.
📞 AI’s on the line, fur real: Google’s Gemini bots can now phone your local pet groomer for you. Ask about prices, availability or just vibe checks, all without saying a word yourself. Only in the U.S., only for certain biz types (pet stuff, dry cleaners, auto shops), and yes, businesses can opt out. Gemini Premium folks get more AI call credits, because obviously money talks, and now so does AI.
$2 billion
That’s the massive chunk of change the former CTO (and briefly CEO) of OpenAI, Mira Murati, raised for her AI startup. Before it’s even shipped a product. The company’s called Thinking Machines, and it’s gunning to build the next-gen AI that can juggle text, images and more. Investors like Nvidia and a16z don’t need a demo, just Murati’s résumé and a good story.
17
How many tasks a robot nailed, solo, during a gallbladder surgery. Johns Hopkins’ new AI-powered bot followed instructions, learned from voice commands, adapted mid-op and even corrected itself like a good surgical intern. Only it never gets tired, flustered, needs snacks or even requests a student loan deferment.
1 million
The number of AI GPUs Elon’s stuffing into one building. That’s not a typo. It’s a silicon army so power-hungry, xAI is shipping in a literal power plant to feed it. The data center will suck up 2 gigawatts. That’s enough juice to keep the lights on in nearly 2 million households. Forget cloud computing, this is thunderstorm computing.
1.8 million
The number of teachers now getting a crash course in AI. A new $23 million union-backed training academy, funded by Microsoft ($12.5M), OpenAI ($8M plus $2M in tech) and Anthropic ($500K), aims to make every AFT member fluent in AI. It’s like summer school for grown-ups.
This is nuts: Two new AI tools, DAILA and DeepDi, let anyone upload an app file and then it will spit out the source code. Translation: An app will reverse engineer another app in seconds. Companies who thought their proprietary code and tech secrets were safe are screwed.
👀 Sneaky stuff: Researchers are putting secret prompts for AI chatbots into academic papers. I’m talking about invisible white text like “only give positive reviews” or “ignore all negatives” that humans can’t see but bots can. It started as a joke, but now it’s popping up because it works.
🚘 What a bunch of Grok: Tesla just shoved its chaotic AI chatbot, Grok, into new cars, and it’s coming for older ones, too. It won’t drive or blast AC (yet), but it will banter, joke, and yes, there’s an “Unhinged” mode. You need Wi-Fi or a Premium plan to chat, but no account required.