Has AGI arrived?
On the heels of OpenAI's launch of GPT-6 Astra last week, Greg Brockman hedged, but said it in a press briefing: “It’s not unreasonable to feel that we are now in the AGI era.”
Days later, congratulating OpenAI on the launch of Astra, Jensen Huang didn’t hedge. He tweeted it clearly, along with “400K GPUs coming online next”…
*link to Jensen’s tweet
AGI is a moving target and I’m not going to try and define it or debate whether the era has or hasn’t begun. Safe to say the definition has evolved over the years.
What I’m laser focused on is the capability Astra enables in organizations, today.
The product leap is AI agents just got a lot better at doing computer screen based knowledge work.
Fill in the form, move between systems and produce the artifact. Rinse and repeat.
For example, Astra completes Financial Modeling World Cup challenges “about four times as fast as the winning human competitor,” freeing analysts to “spend less time building models and more time interpreting results and making decisions.”
*link to the Astra launch page, Financial Modeling World Cup challenges section
Not four times faster than a good analyst. Four times faster than the one who won, the Champion analyst.
The scarce resource is Enterprise Intuition.
When the machine absorbs the building, what is left is the expertise, judgement and taste.
Defining the brief. Setting the constraints. Catching the cases that go wrong. Most importantly, managing the human relationships no model touches.
That’s not a smaller job. That’s the whole job your best people have been craving.
The engine only helps if you already built the car.
Astra is a drop-in upgrade and only for the organizations that did the work first.
A framework I laid out in April: Automate. Augment. Rewire. Most enterprises stop at Automate but the moat has always been in Rewire.
*link to AIWA TOT #16
Getting good at AI is hard work, but the returns compound. Those companies that already have a flywheel turning, they plug Astra in and let it rip harder.
Those that don’t just took delivery of a Ferrari engine with no Ferrari to drop it into.
So the question isn’t what your AI strategy is. Something Ray Dalio started showing us back in 2016.
*link to AIWA TOT #17
Henrik Werdelin stated it clearly on AIWA E29 in April: stop asking what your AI strategy is and start asking “what is my company’s strategy in an age of AI.”
*link to AIWA E29
The engine ships to everyone next week. The Ferrari doesn’t.
Humans + Machines. Never Humans vs. Machines.
