Gary Marcus
Gary Marcus
Contributor, The Yuan

Gary Marcus is a scientist, best-selling author, and serial entrepreneur. He is the founder of Robust.AI and Geometric.AI, acquired by Uber. He is well-known for his challenges to contemporary AI, anticipating many of the current problems decades in advance, and for his research in human language development. An Emeritus Professor of Psychology and Neural Science at NYU, he is the author of five book, and has often contributed to The New Yorker, Wired, and The New York Times.His most recent book, co-authored with Ernest Davis, Rebooting AI, is one of Forbes’s 7 Must Read Books in AI.

Looking ahead, OpenAI’s got 9.9 problems, and Twitch ain’t one
Cognition
While late 2022 and 2023 were heady times for ChatGPT’s creator OpenAI, the AI darling faces many serious challenges in 2024 as its momentum slows and cracks in its business model become more apparent, argues best-selling AI author, entrepreneur, and professor Gary Marcus.
Gary Marcus  |  Apr 23, 2024
Far from being amazing, Sora seems unable to handle the truth
New era
Sora, a text-to-video and text-to-image AI model from OpenAI, is known for creating realistic scenes. A closer look, however, reveals that many of these are not real and should not be mistaken for such, warns Gary Marcus, a best-selling AI author, entrepreneur, and professor.
Gary Marcus  |  Mar 11, 2024
What was 60 Minutes thinking in their recent interview with Geoff Hinton?
Cognition
In an interview with Geoff Hinton, a so-called ‘godfather of AI,’ journalist and CBS news anchor Scott Pelley did not exactly do his homework. Though a missed opportunity, it also offers useful lessons, asserts Gary Marcus, a best-selling AI author, entrepreneur, and professor.
Gary Marcus  |  Feb 06, 2024
Will the desperate race to save generative AI succeed?
Generic
Copyright infringement issues threaten to sink the businesses of GenAI companies like OpenAI and Midjourney. While they do have gambits, these are unlikely to trump underlying concerns, asserts Gary Marcus, a best-selling AI author, entrepreneur, and professor.
Gary Marcus  |  Jan 31, 2024
Open letter: EU’s AI Act must not exclude regulation of foundation models
Governments4AI
The EU lately concluded an agreement on the AI Act, an overarching set of laws crafted to govern AI use within the trade bloc. The following is a letter Prof Gary Marcus and 70 other top global AI experts inked in advance that explains why the Act should encompass foundation models.
Gary Marcus  |  Jan 10, 2024
Without compromise, the EU’s AI Act will unleash sheer chaos
Generic
The EU is at a crossroads with AI. Many have lauded its AI Act as evidence the bloc has a blueprint for sustainable, regulatable AI, but the devil lies in the details as parties disagree over implementation. Best-selling AI author, entrepreneur, and NYU Prof Gary Marcus explains.
Gary Marcus  |  Dec 14, 2023
AI has generated plenty of hype, but what about its economic performance?
Emerging markets
In trying to predict which are likeliest to endure, economic and financial indicators are the tell-tale signs, says best-selling AI author, entrepreneur and NYU Prof Gary Marcus.
Gary Marcus  |  Oct 05, 2023
Doug Lenat (1950-2023) was a true giant in the global AI community
People
Doug Lenat, a prominent computer scientist and AI researcher, recently passed away from cancer. Fellow AI expert Gary Marcus reflects on Lenat’s achievements and legacy, as well as the impact he had on the AI community and what lessons there are for the future.
Gary Marcus  |  Sep 12, 2023
AI risk should not be conflated with AGI risk, though both are dangerous
New era
Superintelligence may or may not be imminent, but great cause for concern exists either way, and one should first understand the differences between AI risk and AGI risk, explains AI expert Prof Gary Marcus, co-host of the podcast Humans versus Machines.
Gary Marcus  |  May 10, 2023
GPT-4 has achieved many successes, yet also has many failings
Cognition
Many are wondering how GPT-4 fits into the larger tapestry of the quest for artificial general intelligence (AGI). To understand this, not only should one know which tasks it performs well, but also its shortcomings and why these are problematic.
Gary Marcus  |  Apr 04, 2023
Does GPT-4 augur the spark of AGI or end of science?
New era
Marching into the future with an obstructed view is a bad idea, yet that seems to be happening with the rollout of GPT-4. For any of this to live up to its potential, there must be more transparency so the scientific community can identify areas of improvement and mitigate risks.
Gary Marcus  |  Mar 29, 2023
Inside the heart of ChatGPT’s darkness: Nightmare on LLM Street
New era
As ChatGPT keeps making waves and more is known about its abilities - and its shortfalls - terrifying possibilities loom, particularly if OpenAI cannot fix its bugs. Still, even if the bot goes down in history as nothing but a publicity stunt it has already redefined possibility.
Gary Marcus  |  Feb 28, 2023
Is Microsoft’s big bet on OpenAI a slam dunk or long shot?
New era
Amid the razzmatazz surrounding ChatGPT, several big players like Microsoft are wagering on this and other promising AI tech. This is risky but not necessarily foolhardy. The coming months will show whether these investments prove wise, says best-selling AI author, entrepreneur, and NYU Prof Gary Marcus.
Gary Marcus  |  Feb 06, 2023
Is ChatGPT really a ‘code red’ for Google Search? Maybe not
New era
ChatGPT has lately been the focus of a great deal of buzz and great expectations, though its real capabilities and limitations also warrant attention, especially when compared to Google’s search engine. AI scientist, best-selling author, and serial entrepreneur Gary Marcus cuts through all the ballyhoo in quest of answers.
Gary Marcus  |  Feb 02, 2023
An epic AI Debate - and why everyone should fret over AI in 2023
Optimization
A time capsule of AI thought leaders participating in a video debate in 2022 serves up a bounty of food for thought that will remain of interest for years to come. AI holds plenty of promise, but major grounds for concern are its less benevolent aspects and the inescapable reality that it still falls far short.
Gary Marcus  |  Jan 24, 2023
What to expect when one is expecting...GPT-4
New era
ChatGPT became a sensation last year for its ability to spit out human-like text in response to questions. This year is almost certain to see the release of GPT-4, which promises to be even more impressive. Will GPT-4 represent the next step on the path to even greater advances, or will it be a major breakthrough in and of itself?
Gary Marcus  |  Jan 09, 2023
What does Meta AI’s diplomacy-winning Cicero mean for AI? (Hint: It’s not all about scaling)
Metaverse
Meta’s new AI, called Cicero, is able to successfully strategize and negotiate with humans in playing the game Diplomacy, and it may have the potential to surpass other AI that play games such as Go and chess. Nevertheless, much remains unknown, and it could be a long time before its implications for AI as a whole are fully understood.
Gary Marcus, Ernest Davis  |  Dec 13, 2022
Alt intelligence is the new science
New era
AI has lost its way, and the world must take a step back and reassess the situation. Too often, the reality of new AI has failed to live up to the hype, and so one must take a closer look at just what exactly AI is capable of, what is still beyond it, and why. Only then will efficient future breakthroughs be possible.
Gary Marcus  |  Jul 25, 2022