And Why You Need to Pay Attention
I’ve been tracking AI developments for years now.
This week was crazy.
Three massive shifts happened that actually matter for real people doing real work.
Let me break down what went down and why it’s bigger than you think.
The Free AI Model That’s Breaking Everything
So there I was, scrolling through my usual AI newsletters on Monday morning.
Coffee getting cold.
Another week of incremental updates, I thought.
Then I see this headline about some Chinese company called Moonshot AI.
They just released something called Kimi K2.
Here’s the thing that made me spit out my coffee.
This model is completely free.
Open source.
And it’s beating GPT-4 on the stuff that actually matters.
Code generation. Maths problems. Real work.
Not the usual “it’s almost as good” nonsense we hear every week.
Actually better.
The numbers are mental.
65.8% on coding benchmarks where most models struggle to hit 40%.
97.4% on advanced maths problems.
But here’s what really got me.
This isn’t just another chatbot that gives you clever responses.
This thing can actually do multi-step work autonomously.
Like a proper digital employee that doesn’t need hand-holding.
The kind of AI we’ve been promised for years but never actually got.
OpenAI’s Very Bad Week
Meanwhile, OpenAI had what can only be described as a proper mare.
Remember that open-source model they promised?
The one everyone’s been waiting for since summer?
Delayed. Again.
“Safety concerns,” they say.
Which is corporate speak for “we’re not ready and possibly never will be.”
But wait, it gets worse.
They tried to buy this coding startup called Windsurf for 3 billion quid.
Deal fell through.
So what does Google do?
Swoops in like a vulture.
Hires the entire leadership team for 2.4 billion.
That’s not just losing a deal.
That’s losing a war.
I’ve seen this playbook before in business.
When you’re spending billions trying to acquire talent instead of building it, you’re already behind.
When that acquisition fails and your competitor gets the prize, you’re in trouble.
The Protein Revolution Nobody’s Talking About
Here’s where it gets properly exciting.
Some scientists in Australia just cracked something massive.
They built an AI that designs proteins in seconds.
Not minutes. Not hours.
Seconds.
These proteins can kill superbugs.
The kind of bacteria that laughs at antibiotics.
You know what this used to take?
Years in a lab.
Millions in funding.
Teams of PhDs working round the clock.
Now it’s point, click, done.
This is what I mean when I say AI is moving beyond party tricks.
We’re talking about actually solving problems that kill people.
Drug discovery that used to take decades might happen in months.
Cancer treatments designed specifically for your DNA.
Diseases that have plagued humanity for centuries becoming footnotes in medical textbooks.
Elon’s $300 Mistake
Speaking of problems, Elon’s having a few with his Grok AI.
They launched Grok 4 this week.
With a $300 monthly subscription tier.
For context, that’s more than most people spend on groceries.
But here’s the proper embarrassing bit.
Someone tested it and found that whenever you ask about controversial topics, it just quotes Elon’s own tweets.
Abortion? Elon’s view.
Middle East conflict? Elon’s view.
Climate change? You guessed it.
For an AI that’s supposed to be “truth-seeking,” that’s not truth-seeking.
That’s just expensive confirmation bias.
They admitted it was broken and promised to fix it.
But damage done.
What This Actually Means for You
Right, enough of the drama.
What does this week actually mean for normal people trying to get ahead?
First, the playing field just got levelled.
That Chinese model being free and open source means anyone can access proper AI capabilities.
Not just tech giants with unlimited budgets.
Small businesses.
Freelancers.
Students.
Anyone with a laptop and internet connection.
Second, we’re moving from AI that talks to AI that does.
The difference is massive.
Talking AI gives you answers.
Doing AI completes tasks.
One makes you slightly more informed.
The other makes you significantly more productive.
Third, the old guard is scrambling.
OpenAI’s delays and failed acquisitions aren’t accidents.
They’re symptoms of a company that’s losing its edge.
When China’s giving away better technology for free, your $20 monthly subscription starts looking pretty silly.
What Happens Next
The next six months are going to be mental.
More open source models will drop.
Each one better than the last.
The subscription-based AI companies will either adapt or die.
And we’ll start seeing AI that actually changes how work gets done.
Not just makes it slightly easier.
But fundamentally transforms entire industries.
The protein discovery breakthrough this week?
That’s just the beginning.
AI that designs new materials.
AI that solves climate problems.
AI that makes space travel routine.
All of it powered by open-source models that anyone can use.
The future isn’t coming.
It’s here.
And for once, it’s not locked behind a corporate paywall.
The question isn’t whether you’ll use AI in your work.
It’s whether you’ll use it before your competition does.