Gary Marcus is now one of the loudest skeptics of the AI boom. In 2012, almost nobody was listening.
I have the tape.
That year, I sat down with him for Singularity. FM, right after he published a sharp critique of Ray Kurzweil’s theory of mind in The New Yorker. Marcus was already making the argument that would define his career. Intelligence is not just pattern-matching. The mind is a kluge, a messy evolutionary patch job. And scale alone will not get you to real #AI.
More than a decade later, that argument is everywhere. Labs are chasing the hybrid and neurosymbolic approaches he pointed to back then. The field finally caught up to the conversation.
But here is what makes the interview worth revisiting. He also bet big on neuroscience as the road forward, on projects like Blue Brain and Whole Brain Emulation. The breakthroughs came from somewhere else entirely.
So was he the prophet, or just early on some calls and wrong on others? Watch it and decide for yourself.
Douglas Hofstadter, the Pulitzer Prize–winning author of Gödel, Escher, Bach, voices his concerns about how the current wave of rapid advancements in AI may endanger humanity.
CHAPTERS 0:00 Introduction. 0:34 When I started out, computers were rigid. 1:29 I thought Artificial Intelligence would take hundreds of years. 1:59 I never imagined computers would rival humans so soon. 2:53 It feels like humans are about to be eclipsed. 4:01 I feel diminished, inferior. 5:01 AI pioneer Geoff Hinton may regret part of his life’s work. 6:07 Conclusion: what do you think?
WATCH THE FULL INTERVIEW • Gödel, Escher, Bach author Doug Hofstadter…
Researchers at the Korea Advanced Institute of Science and Technology (KAIST) have developed a technique to carve microscopic liquid-cooling channels directly inside silicon semiconductor chips.
Interestingly, the computer architecture slashed the energy required for cooling by pumping ordinary, room-temperature water straight through the chip’s internal structure.
“As the performance of AI semiconductors and advanced electronic packaging becomes increasingly limited by heat, we expect this technology to serve as a foundational cooling solution for future high-performance computing systems,” said Professor Sung Jin Kim.
The Path to Robust deAGI asks what it would take to build artificial general intelligence that is both powerful and structurally aligned with human flourishing—not just steered by after‑the‑fact safety patches. Ben Goertzel, CEO of SingularityNET and a founding member of the Artificial Superintelligence (ASI) Alliance, will outline how a decentralized, token‑coordinated ecosystem—combining ASI: Chain, Hyperon AGI, and community‑owned GPU clouds—can prevent AGI from being captured by any single corporation or state.
Goertzel will contrast centralized AGI roadmaps with a deAGI approach that bakes openness, diversity of values, and economic inclusion into the architecture itself, drawing on ideas like pluralistic training data, interoperable agent networks, and on‑chain governance of key system upgrades. He will also discuss technical milestones toward “robust” deAGI—modular cognitive architectures, decentralized marketplaces for AI services, and verification mechanisms that let communities audit and constrain AGI behavior—framing them as concrete steps toward an AGI that advances joy, growth, and choice for all rather than amplifying existing power imbalances.
To accelerate and refine decision-making in a fast-paced, global marketplace, enterprises may deploy generative artificial intelligence models to help summarize and interpret the charts that often fill market summaries and financial reports.
But even the latest vision-language models sometimes struggle with this task, since it requires a model to integrate visual, numerical, and linguistic understanding. A company that invests in a state-of-the-art model might still receive inaccurate or incomplete information.
To fill this performance gap, researchers from MIT and the MIT-IBM Computing Research Lab developed a multifaceted resource for AI users that is specifically designed to teach vision-language models (VLMs) how to effectively interpret charts.
Flocking birds and schools of fish are a familiar sight. While previous research has uncovered the broad dynamics driving these movements, their underlying intricacies remain a mystery. Now a study by a team of New York University mathematicians offers new insights into these phenomena. It reveals that flocks and schools behave in ways similar to a soft crystalline material, with individual birds and fish serving as “atoms” that are evenly spaced in a lattice-like formation.
The findings, reported in the journal Physical Review Fluids, offer detailed insights into the hydrodynamic and aerodynamic interactions crucial in aerospace and automotive engineering, robotics and energy harvesting.
“Our findings offer a new way to understand how animal collectives coordinate movement and respond to their environment,” says Christiana Mavroyiakoumou, a researcher at NYU’s Courant Institute School of Mathematics, Computing, and Data Science at the time of the study and now a fellow at Oxford University’s Mathematical Institute. “More specifically, lines of birds or fish behave like an elastic material with regularly spaced individuals held together by flexible, or spring-like, bonds—akin to soft crystalline substances in which atoms are arranged in an orderly, repeating pattern.”
