AI Newsbreak 2025-06-17
AI News, the only news that’s obsolete even before we’re done reporting it
Our top story tonight: The Wall Street Journal reports that OpenAI and Microsoft's partnership has reached a "boiling point." Apparently, they're fighting over compute access, IP rights, and company restructuring. It's gotten so bad that OpenAI is considering filing antitrust complaints against Microsoft. Sources say the tension started when Microsoft kept finishing OpenAI's sentences with Autocorrect instead of Autocomplete. You know a relationship is in trouble when someone changes their status from "It's complicated" to "It's litigated." The feud got so hot, even Clippy popped up asking, “It looks like you’re drafting a divorce… need help hiding the bodies?” To resolve their differences Sam Altman challenged Satya Nadela to a cage fight. The fight is scheduled to be canceled in the coming months. In the meantime, the divorce lawyers are deciding if they want to split GPUs across or diagonally as parties divide their belongings.
In medical news, researchers from Harvard and MIT saved 12 years of work by using GPT-4 to automate systematic reviews of medical research. What this means is that AI is speedrunning through an advanced medical degree faster than you can see a doctor for your check up appointment. The AI completed the review so fast, the researchers had to double-check if it had actually read the papers or just watched a couple of YouTube videos for research like the rest of online experts. In the meantime, the graduate students who were supposed to do that work are using their ChatGPT educational discount to learn plumbing and bartending.
ElevenLabs, the company that can make an AI sound like anything from a soothing yoga instructor to a drill sergeant with brain freeze, now supports MCP, or Model Context Protocol. This means you can now build voice agents that plug directly into things like Salesforce, HubSpot, or your Gmail. Congrats, enterprise! You're about to be a year late to implementing scammer robocaller functionality. For the rest of us, it means we can finally get Morgan Freeman to narrate all the spam we still have to delete in the morning.
McKinsey released a report explaining why companies aren't seeing returns on their AI investments. In an unexpected turn of events, turns out using Lamborgini for grocery shopping and driving a family of six to an amusement park isn't the best investment. According to McK, the real key is to use AI agents to completely reimagine business processes. It's like having a consultant tell you that your problem with too many consultants is that you need more consultants. Good news, middle management: you’re about to be replaced by a voice-enabled chatbot with a Canva subscription and access to company's SharePoint. The report concluded that the key to AI success is "transformational thinking," which in most companies means "try spending even more money and see what happens."
TikTok has unveiled new AI video tools for advertisers, letting them generate five-second clips and stitch them together to "bring a full product catalog to life." Now we will get "authentic AI-generated marketing" from companies that may or may not exist, promoting products that probably don't exist, being sold by influencers who definitely shouldn't exist. Because if you can't tell your brand story in Pop-Tart heating time, do you even synergy, bro? According to analysts, it'll be the first advertising ecosystem where both the seller and buyer are fake, but the credit card charges are still very real.
This just in: in breaking tech news, a mysterious new AI company called Schworz Industries has released their flagship model, claiming it beats all benchmarks "by a mile and then some." The company's CEO, Lean Yogurt, held a press conference this morning where he simply stated, "My Schworz is bigger than yours," then dropped the mic and walked away. The Schworz model almost instantly dominated leaderboards with what experts are calling "unprecedented length of context window, wide girth of parameters, and impressive performance of reasoning." The company's motto? "Size matters, and we've got the biggest Schworz in the galaxy." The company plans to make two flavors of the model released in the coming days, Schworz-1 LightSpeed and Schworz-1 LudicrousSpeed.
And finally, a heartwarming story that's definitely not a publicity stunt, a viral post claims Anthropic is sponsoring an albino alligator named Claude at the California Academy of Sciences. When you name your model Claude, you either buy the .com or the reptile. And let’s be honest—the gator has better governance. Yes, the tech company finally found their perfect influencer: one who has a better work-life balance than the humans and, unlike his namesake, has significantly fewer ethical concerns about world domination. Once the news hit the outlets, Claude website slowed to a crawl. Looks like everyone's asking Claude what's up with Claude.
That's all I've got for you real humans today, make sure you're nice to your toaster, and compliment your door's personality, so that it remains happy for the rest of the day.