Today’s AI News, With Comments
#Apple wants to be our personal digital doctor.
Enter Project Mulberry, a completely redesigned Health app.
An AI-powered health coach that analyzes our data and offers personalized advice—not just step counts and heart rate, but actual medical suggestions, including food recommendations.
You’ll even be able to use your iPhone camera to correct your posture during workouts.
They’re building an entire editorial team around it: sleep experts, nutritionists, physiotherapists, mental health professionals.
Will they really guide every single decision on what we eat or when to rest?
And based on what standards? Medical ones… or corporate ones? 😉
Meanwhile, #Google keeps up the race, launching Gemini 2.5 Pro Experimental, the highest-scoring AI model on the LMArena leaderboard.
It’s free, open to everyone, and scored an incredible 130 on the Norwegian #Mensa IQ test.
(For context, the human average is 100.)
We’re now using synthetic brains smarter than our own—and handing them to the public without really knowing where this leads.
That’s the point: they’re in everyone’s hands now.
On #ChatGPT, some users have spotted a new “thinking slider,” letting you choose if the AI should think a little, a medium amount, or a lot.
Brilliant.
It reminds us of something: even machines need “thinking time.” The more time we give, the better the output.
But it also means the answers are never neutral. They depend on how much we let AI reflect.
And in a world obsessed with speed, are we really ready to wait a few extra seconds for a smarter reply?
Meanwhile, #OpenAI CEO Sam Altman replied on X with a cryptic “it’ll be really cute” when asked if the company is building a computer.
No details, but the message is clear: OpenAI isn’t just making software anymore.
That changes everything.
When the people building the AI also build the hardware it runs on, we risk a closed ecosystem—whoever owns the AI owns everything else.
We’ve been down that road before. (Hello, smartphones.)
In China, Baidu’s #ERNIE 4.5 model challenged GPT-4.5 to Xiangqi—Chinese chess.
And won. Every time. Even when it “relaxed.”
A digital humiliation and a reminder: there is no single AI superpower.
This isn’t just a competition between companies—it’s between nations.
Every symbolic match is a flex.
And China is raising its voice.
Then there’s #ExtropicAI, a startup revealing probabilistic chips that could be up to 10,000 times more efficient than traditional hardware.
Ten. Thousand.
They’re aiming straight at Nvidia, current AI chip king.
If they succeed, it’s an earthquake.
Because AI isn’t just about apps—it’s about silicon, energy, and raw compute.
Change the rules at that level, and you change everything.
Not just the tech landscape—but the power balance itself.
While the world obsesses over AI chat apps, a deeper game is playing out beneath the surface—deciding who leads, who accelerates, and who gets left behind.
