News

Microsoft AI Chief Predicts AGI Still "About a Decade" Away, Contradicting Accelerated Timelines

End Of Miles

06 Apr 2025 — 2 min read

"It doesn't instinctively feel to me like we're 2 to 3 years away. I know some people think that it is and I respect them deeply. I feel like we're still a good decade or so away," said Microsoft AI CEO Mustafa Suleyman, offering a notably more conservative timeline for artificial general intelligence than many of his industry counterparts.

End of Miles reports that Suleyman's assessment stands in contrast to accelerating timelines proposed by other leading AI labs, including his former colleague Demis Hassabis at Google DeepMind, who has suggested AGI might arrive within 3-5 years.

A measured perspective amid the AGI race

The Microsoft executive acknowledged the difficulty in predicting exact timelines for such a transformative technology. "When a scientist or a technologist or an entrepreneur like me says we're a decade away, that's just a hand-wavy way of saying we're not really sure and it feels pretty far off," Suleyman explained.

"I could imagine it happening within 5 years. Absolutely. The rate of progress over the last 3 or 4 years has been electric. It's kind of unlike any other explosion of technology we've ever seen." Mustafa Suleyman, CEO of Microsoft AI

Despite recognizing the possibility of earlier breakthroughs, he emphasized that significant technical challenges remain before achieving true AGI. "Instinctively to me it feels like there's still a lot of basics that we got to get right," he said.

Technical hurdles still to overcome

Suleyman identified several specific areas where current AI systems fall short of AGI capabilities. "We still have to nail hallucinations. We still have to nail citations. It's still not great at instruction following. It still doesn't quite do memory. It still doesn't personalize to every individual," the AI leader explained.

While acknowledging these limitations, he also noted promising developments: "We're seeing the glimmers of it doing all of those things. We're taking steady steps on the way there."

"If we really are on the cusp of producing something that is more valuable than all economically productive work that any human can produce, I think one of the last things we're going to be worried about is our partnership with OpenAI. It's going to profoundly change humanity." Suleyman on the potential impact of AGI

Reframing the AGI conversation

The Microsoft AI CEO suggested an alternative framework for discussing artificial intelligence development. "Everyone gets a little bit caught up on week-to-week, day-to-day, or definitions of these abstract ideas. Just focus on the capabilities," he advised.

Suleyman proposed "artificial capable intelligence" as a more practical concept than AGI. "It's more measurable and we can actually look at it very explicitly in terms of its economic impact and its impact on work," he explained.

This perspective aligns with his broader prediction of a 15-year transition where AI agents will fundamentally transform knowledge work, suggesting that practical AI capabilities may have profound economic impacts well before achieving technical definitions of AGI.

Advanced AI processor contained in security field with prismatic verification protocols; AI chip regulation; compute security framework; technological safeguards

Tech Leaders Propose Treating AI Chips Like Nuclear Materials

Former Google CEO Eric Schmidt and co-authors have proposed implementing controls for advanced AI chips similar to those used for nuclear materials, arguing that compute resources represent a critical national security chokepoint that demands unprecedented oversight. The group's recommendations include geolocation tracking, firmware-level security, and strict export controls

Neural network with autonomous data streams glowing in iridescent spectrum; visualization of AI systems learning through experience beyond human limitations

Google DeepMind Researchers Propose "Streams" as Path to Superintelligence

"Experiential data will eclipse the scale and quality of human-generated data," declare legendary AI researchers David Silver and Richard Sutton in a groundbreaking new paper that outlines how autonomous AI systems could develop capabilities "that surpass those possessed by any human." The two Google DeepMind luminaries

Students Engage With AI in Four Distinct Patterns, Anthropic Education Report Reveals

University students are using AI systems like Claude in four distinctly different patterns, each appearing at surprisingly similar rates in student conversations, according to groundbreaking research analyzing real student interactions. End of Miles reports that these interaction patterns challenge the common perception that students primarily use AI just to find

Cellular gene expression transforming into language sequences with iridescent gradients; C2S-Scale LLM technology bridges biology and AI

Google and Yale Create 'Virtual Cells' That Could Replace Lab Experiments

"One of the most exciting applications is forecasting how a cell will respond to a perturbation — like a drug, a gene knockout, or exposure to a cytokine," reveals Yale Assistant Professor David van Dijk about a breakthrough AI system that can simulate cellular behavior without traditional lab experiments.