OpenAI’s Panic Response to Gemini

In partnership with

Hey there,

For the first time in a while, OpenAI isn’t at the center because of a new release or update, but because of panic.

Sam Altman recently sent a company-wide memo declaring “code red” for ChatGPT.

The reason? Google is catching up. Fast.

Google’s latest Gemini models and its growing AI ecosystem are putting serious pressure on OpenAI’s lead.

People have started using Gemini more; you can see the shift in usage trends here.

OpenAI has decided to pause projects like ads, shopping agents, health agents, and even its personal assistant “Pulse.” Now they’re just focusing on one thing:

Improve ChatGPT. Make it faster, more reliable, and more helpful.

Honestly, it was about time. Competition forces companies to wake up. And this kind of healthy pressure usually results in better products for all of us.

Microsoft Is Struggling Too

Microsoft, OpenAI’s biggest partner, isn’t having a great time either.

The company expected businesses to buy a lot of its new AI “agents,” but most customers didn’t. 

Now Microsoft has cut its sales targets because salespeople couldn’t meet their goals.

Why?

Because the tech still feels early:

• Agents make mistakes

• They break on unfamiliar tasks

• Businesses don’t trust them for high-stakes workflows

Microsoft had pitched 2025 as “the era of AI agents.” But right now, the era isn’t selling. 

It’s not that agents themselves are the problem; it’s the education gap. Most businesses still don’t understand how to integrate AI into real operations. They want the outcome but don’t have the processes, data, or trust in place yet.

Open Source Models Keep Getting Better

While OpenAI and Microsoft deal with internal fires, Mistral quietly dropped its biggest release yet: Mistral 3.

And honestly… it’s impressive.

They launched:

• Huge open-source model (Mistral Large 3)

• Three small, insanely efficient models (3B, 8B, 14B)

• All multimodal

• All multilingual

• All open weights under Apache 2.0

The small models especially are a problem for big players:
They can run on laptops, edge devices, and even consumer GPUs, and outperform models way larger than them.

In short: while the giants are fighting each other, Mistral is arming the entire developer ecosystem.

What Happens When the “AI Bubble” Pops?

The Guardian published a deeper take that’s worth noting.

They argue the real issue isn’t if the AI bubble will burst, it’s what it will leave behind.

Every major technological rush has had a bubble: railroads, the internet, crypto and even tulips (yep, tulips).

Some bubbles leave long-term value (like the internet). Others leave nothing but a crash (like subprime mortgages).

The AI bubble is different because:

• Big Tech is taking on massive debt

• Money is flowing in circles

• No one knows if LLMs are actually the right path to AGI

Researchers like Yann LeCun now openly question if today’s LLM-based systems can ever reach true intelligence, meaning trillions may be invested in the wrong direction.

So Where Are We Right Now?

We’re watching three realities collide:

1. OpenAI is no longer unchallenged.
   Google’s momentum is forcing them into emergency mode.
   

2. Microsoft’s bet on AI agents isn’t landing yet.
   Enterprises aren’t ready for full autonomy.
   

3. Mistral is quietly becoming the open-source backbone of AI.
   Their models are getting good, fast.

And somewhere underneath all this… the AI bubble might be inflating faster than we think.

What This All Means for You

If you’re simply “using AI,” that won’t cut it anymore. This next phase requires you to become strategic, not just enthusiastic.

AI is no longer about finding the “best” model. There won’t be one best model. Things are moving too fast, and even OpenAI is showing cracks. The real advantage now comes from your ability to adapt as things shift.

If you’re a business leader, it’s time to get clearer about what AI is for. Not hype. Not “agents that will run the company for you.” 

Start with the boring stuff, the workflows that waste hours, the decisions that repeat daily, the customer problems that don’t go away. 

Real ROI in AI right now isn’t coming from autonomy. It’s coming from efficiency. 

PS: I’m curious what you think about all this. Just hit reply, I’d love to hear your take.

- Aashish

200+ AI Side Hustles to Start Right Now

AI isn't just changing business—it's creating entirely new income opportunities. The Hustle's guide features 200+ ways to make money with AI, from beginner-friendly gigs to advanced ventures. Each comes with realistic income projections and resource requirements. Join 1.5M professionals getting daily insights on emerging tech and business opportunities.

Get the guide