No GPT-5 in 2025 and No AGI — Ever

The Triadic Nature of Meaning-Making and the Fallacy of AI’s Understanding

Workshop of Philip Galle after Jan van der Straet, Impressio Liborum (Printing Books), c. 1590/1593. Engraving on laid paper. Plate: 20.1 × 26.5 cm (7 15/16 × 10 7/16 in.), sheet: 27 × 36 cm (10 5/8 × 14 3/16 in.). Nova Reperta (New Discoveries) [plate 4]. Rosenwald Collection. Public domain. Courtesy of the National Gallery of Art, Accession Number 1964.8.1578

Not a member? Read this for free through my friend link.

Scientist and popular AI commentator Gary Marcus recently published¹ a note on his Substack arguing that Artificial General Intelligence (AGI) is not arriving in 2025 — and neither is OpenAI’s much-anticipated GPT-5. Citing insider remarks from a recent conference, Marcus highlights that while GPT-4.5 may be on the horizon, OpenAI’s progress toward GPT-5 appears stalled, reflecting the broader struggles with scaling meaningful breakthroughs in AI.

At a conference yesterday, someone with very good knowledge of OpenAI (Chatham House rules forbid me from naming the source, but allow me to repeat the remark)) said something fascinating, more for what was not said than what was said. What was said is that we should expect to see GPT 4.5 soon. What was not said by the well informed source was that we should expect GPT-5 anytime soon. — Gary Marcus

Marcus attributes these delays to bottlenecks in data and the inherent limitations of Large Language Models (LLMs). He emphasizes that merely scaling data is no longer enough to deliver robust AI capable of approaching human cognition. He also (humorously) notes a shift in Elon Musk’s rhetoric…