You might have missed it if you were immersed in holiday merriment. But a growing chorus of AI experts, from Anthropic engineers to Google researchers to the founder of Midjourney, noticed that in late December, the latest AI models entered incredible new territory.
Claude Opus 4.5 and its accompanying Claude Code tool are getting renewed attention and buzz as we enter 2026 for their perceived leap in capabilities. Some AI watchers have gone so far as to call them the closest thing we have to AGI.
In fact, the conversation around these tools has been so intense that some experts believe we've hit a tipping point in AI capabilities.
To understand what's going on here, I turned to SmarterX and Marketing AI Institute founder and CEO Paul Roetzer on Episode 189 of The Artificial Intelligence Show.
A Fast-Paced Timeline
The buzz started quietly the day after Christmas when Igor Babuschkin, co-founder of Elon Musk's xAI and former researcher at Google DeepMind and OpenAI, posted a simple observation on X: "Opus 4.5 is pretty good."
Andrej Karpathy, co-founder of OpenAI and one of the most respected voices in AI research, replied with something more pointed: "It's very good. People who aren't keeping up even over the last 30 days already have a deprecated worldview on this topic."
That same day, Jackson Kernion, an Anthropic employee of four-plus years, dropped a bombshell that would ripple across tech circles. He wrote that he joined Anthropic motivated by the dream of building AGI, and now feels like "Opus 4.5 is as much AGI as I ever hoped for."
The reactions to his pronouncement were swift and polarizing. But Kernion doubled down the next day, noting that skeptics simply hadn't experienced what Claude Code could do:
"To use Claude Code is to see Claude write arbitrary software, run into errors, reliably fix them, make helpful suggestions, and perfectly follow any given instructions."
By January 2, the conversation had exploded. Jaana Dogan, a principal engineer at Google, posted what would become a viral thread with millions of views: "I'm not joking and this isn't funny. We have been trying to build distributed agent orchestrators at Google since last year. There are various options, not everyone is aligned ... I gave Claude Code a description of the problem, it generated what we built last year in an hour."
The next day, David Holz, founder of Midjourney, added: "I've done more personal coding projects over Christmas break than I have in the last 10 years. It's crazy. I can sense the limitations, but I *know* nothing is going to be the same anymore."
And Elon Musk responded to Holz's post with some simple, powerful words:
"We have entered the singularity."
The Data Behind the Hype
This isn't just tech insiders getting excited. Hard data supports their enthusiasm.
The evaluation group METR estimated that Claude Opus 4.5 has what they call a "time horizon" of nearly five hours. This measures how long AI can perform a task autonomously with a 50% success rate, relative to how long a human would take to complete it. This is the highest time horizon METR has ever published, and researchers noted their current testing suite is actually reaching its limits for measuring these upper bounds.
Meanwhile, Epoch AI's analysis indicates that the overall rate of AI progress has nearly doubled in the last two years. Their capabilities index shows a sharp improvement in early 2024, driven largely by the emergence of reasoning models and reinforcement learning techniques.
METR's data specifically shows task horizons are doubling roughly every seven months. If these trends hold, researchers predict AI agents could reliably handle week-long tasks within the next two to four years.
"The labs have smarter, more powerful versions of the models than you do," says Roetzer. "They are seeing and experiencing the world differently than you are. They have access to things that you and I do not."
Amodei’s Prediction and Why It Matters
To understand the significance of what's happening, it helps to rewind to March 2025. Anthropic CEO Dario Amodei gave an interview at the Council on Foreign Relations where he made a bold prediction that many dismissed as hyperbole.
“What we are finding is we are not far from the world — I think we'll be there in three to six months — where AI is writing 90% of the code," Amodei said. "And then in 12 months we may be in a world where AI is writing essentially all of the code."
Roetzer points out that Amodei doesn't tend to exaggerate.
"Usually if he says something like that, he has already seen it or sees a very clear path to something like that occurring," Roetzer says.
Amodei’s prediction now appears to be coming true. And the implications extend far beyond software development.
"The precursor to the disruption of all knowledge work is coding and AI research," says Roetzer, "because that is the most valuable thing for these labs to build.
"So they're fine-tuning models to be great at doing this. If 90% of all code is being written by the AI, that enables them to generate more code, take more shots on goal from an AI research perspective. This is the first domino."
A CEO's Personal Experience
The state of AI isn't just a story about automating research and coding. To illustrate this, Roetzer shared a personal experience.
His company, SmarterX, is at a turning point, with what Roetzer describes as two potential paths: "scale" or "hyperscale." Both involve complex decisions spanning finance, legal, HR, operations, and strategy. These are the kind of decisions that would typically require months of research, dozens of meetings with attorneys and advisors, and countless hours of manual work.
Instead, Roetzer turned to a custom GPT he built called "Co-CEO,” a ChatGPT tool trained on information about his company, revenue plans, growth strategies, and values.
"I basically said, “Here's path A, scale. Path B, hyperscale.’ Let's analyze this together," Roetzer explains.
He told Co-CEO: "As a starting point, create a list of all the questions I should be asking when comparing the two paths. Then we will go through each of those questions one at a time."
Within a minute, ChatGPT generated 56 questions across 10 categories: From North star and personal objectives to market and competitive dynamics to execution readiness.
Roetzer estimates he spent 10 to 15 hours working through this thread. Without AI, he estimates it would have taken him more than 200 hours of manual research, note-taking, summarization, and writing to get to the answers and analysis. The conversation generated more than 20,000 words, nearly half the length of the book he wrote on AI, which he used to create six documents.
“Once I edit and finalize those documents, Co-CEO will have written 95% plus of the final words,’’ Roetzer says.
Humans Still Matter, But In a Different Way
Roetzer is quick to point out that this isn't about letting AI do all the work. Collaboration is key.
"ChatGPT would've never gotten to the end game without me," he says.
"But in the end, the final output is 95% plus what ChatGPT wrote, not words I personally wrote. And I don't see anything wrong with that from a leadership perspective."
It’s a strategic partnership: AI accelerates getting to the point where human judgment, relationships, and expertise actually matter.
"I couldn't call my attorney during break or I couldn't call my banker at random times when I was thinking of stuff," Roetzer explains.
"But if I didn't have this tool, I'd be here on January 5 reaching back out to my attorney saying, 'Can you re-explain that concept?' I don't need that anymore. I did my own work on it. I understand the concept now. Now I can give my attorney direction."
The goal isn't to eliminate the human part. It's to get us there faster.
"I don't want to spend three months figuring out who to hire and job descriptions and all that," says Roetzer. "I want to spend three months interviewing people face-to-face and meeting these people and finding the right people. I want to get to the human part of this. I don't want to spend three months on stuff that the AI is just really good at."
The Bottom Line
What happened over the holidays was a genuine shift in how AI insiders, people building these systems and using them at the frontier, perceive and measure current capabilities.
For most business leaders and knowledge workers, the implications are just beginning to sink in. As Karpathy noted, anyone who hasn't kept up in the last 30 days already has an outdated understanding of what's possible.
"The business world and the future of work has fundamentally changed already," says Roetzer. "Most practitioners and leaders aren't aware of how much."
The coding revolution is just the beginning. Next, the advancements will spread to every form of knowledge work.
"I think the same thing that we're seeing in coding probably happens in business if you know how to use these tools,’’ Roetzer says.
Mike Kaput
Mike Kaput is the Chief Content Officer at SmarterX and a leading voice on the application of AI in business. He is the co-author of Marketing Artificial Intelligence and co-host of The Artificial Intelligence Show podcast.