OpenAI just dropped two game-changing AI models—o3 and o4-mini—and if you’re paying attention, you can feel the seismic shift happening under our feet.
Both models raise the bar for what’s possible with AI today. o3, in particular, isn’t just better at tasks like math, coding, and writing. It’s now capable of reasoning about when and how to use external tools—like searching the web, running code, analyzing images, and generating visuals—without needing endless prompting or manual tool selection. It’s as if you’re working with a highly capable assistant who not only knows what tools to use but when to use them.
And the result? Some serious chatter that o3 might actually qualify as early-stage artificial general intelligence (AGI).
On Episode 145 of The Artificial Intelligence Show, I spoke to Marketing AI Institute founder and CEO Paul Roetzer to get the scoop on o3’s incredible capabilities.
Are We Already Seeing the First Glimpses of AGI?
o3 is shattering records on academic benchmarks and real-world tasks. High-profile figures like economist Tyler Cowen have openly said they believe this model is AGI, writing:
“I think it is AGI, seriously. Try asking it lots of questions, and then ask yourself: just how much smarter was I expecting AGI to be?
As I’ve argued in the past, AGI, however you define it, is not much of a social event per se. It still will take us a long time to use it properly. I do not expect securities prices to move significantly (that AI is progressing rapidly already is priced in, and I doubt if the market cares about “April 16th” per se).
Benchmarks, benchmarks, blah blah blah. Maybe AGI is like porn — I know it when I see it.
And I’ve seen it.”
AI leaders like Scale AI CEO Alexander Wang and former OpenAI Chief Research Officer Bob McGrew are taking notice, too, if not fully committing to o3 being AGI.
Wang calls o3 a “genuine meaningful step forward” thanks to its emergent “agentic” tool use, where it intelligently decides when and how to use external capabilities—an approach powered by reinforcement learning.
McGrew reframes the AGI conversation entirely, saying, “The defining question for AGI isn’t ‘how smart is it’ but ‘what fraction of economically valuable work can it do?'” With o3, intelligence is no longer the primary bottleneck. Instead, it’s about reliable interaction with the external world.
Roetzer isn’t sure we’re at full AGI just yet. But that’s not even the point. The point is: It may not even matter whether or not o3 is AGI.
“I think it’s really important that people continue to remember we don’t need to reach it or agree on it for it to transform everything,” he says.
Not to mention, a more powerful o3 Pro version is reportedly on the way, promising even greater leaps.
(Though, a word of caution: hallucination rates seem to be higher with o3, according to early reports. Roetzer emphasizes vigilance, particularly when using the model for public-facing or high-stakes work.”
Real-World Proof: How o3 Is Already Transforming Work
For evidence of what Roetzer’s talking about, check out these firsthand examples he shared about how o3 is already disrupting knowledge work in ways that are hard to overstate.
On a recent trip to Aruba, Roetzer needed to make a quick but critical decision about upgrading Marketing AI Institute’s office internet to prepare for new employees starting next week—something far outside his expertise. Rather than waiting hours (or days) for IT consultants, like he would have had to do in the past, Roetzer turned to o3. Acting as a senior IT advisor, the model guided him through nuanced technical decisions in real time.
“It helped me understand more deeply how to solve this than any IT person I’ve ever talked to,” Roetzer says. In just 20 minutes, he made a confident, well-informed decision, which saved time, money, and massive headaches.
The story doesn’t stop there. Roetzer also used o3 to work on a complex organizational design project for his company, something that would normally cost $50,000 to $100,000 in outside consulting fees. Instead of receiving a static report from a consultant, he actively engaged with o3, asking questions, challenging assumptions, and iteratively refining the outputs.
Crucially, he plans to vet his final plan by feeding it into other models like Gemini 2.5 for critical evaluation, ensuring even greater confidence.
“You start to increasingly see it doing the things that I would otherwise be paying advisors and consultants to do, or the things that we would traditionally be hiring someone to do,” he says.
“Rather than paying someone to give me a report and say, here’s what you should do, that I would then have to sit there for hours reviewing, analyzing, trying to make sure I understood the recommendations so that I could then make an educated decision. I just did all the work myself with o3.”
A Blunt Wake-Up Call for Professional Services
If you’re in professional services—law, accounting, IT consulting—Roetzer has a blunt message: Run, don’t walk, to spend $200 on unlimited access to o3. Put it through the paces. Test it against the hard questions clients ask you. Because your clients soon will.
“Every time you put a proposal together, you need to be asking yourself, can o3 do this? Could they just use o3 to do this or 80% of this? Because the answer is going to increasingly be ‘yes,'” he says.
Today, only early adopters are thinking this way. But widespread awareness is coming fast. If your offering can be replicated—or at least started—by a capable AI model for a fraction of the cost, expect clients to think twice about hiring you.
The future isn’t waiting. o3 shows that even without “official” AGI, the very fabric of how work gets done is already being rewritten, says Roetzer.
“You can run a business or a department or a team or a campaign in entirely different ways when you know how to work with these tools.”
#OpenAIs #Model #Closest #Weve #AGI
