Quick wins for a faster PC:
Repair Windows errors before they cause bigger problemsFix Now →Scan for outdated or missing drivers - takes under a minuteDriver Scan →Clear out junk files and repair common Windows errorsFree Scan →iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
Gary Marcus argues that today’s large AI systems can produce fluent language without reliably understanding the world, reasoning about cause and effect, or using common sense. He is a cognitive scientist, author, entrepreneur, and prominent AI commentator who favors combining machine learning with structured knowledge and verification—and who also calls for stronger public oversight of AI’s risks.
Who is Gary Marcus?
Gary Marcus is a cognitive scientist whose work bridges human cognition and artificial intelligence. His academic background includes cognitive science at Hampshire College and MIT, and he taught psychology and neural science at New York University, where he is now a professor emeritus. Institutional profiles describe his publications as spanning psychology, neuroscience, genetics, linguistics, and AI.
Marcus is also an entrepreneur. He founded and led Geometric Intelligence, a machine-learning company acquired by Uber in 2016, and later co-founded Robust.AI, a robotics company. Carnegie Council describes him as Robust.AI’s founder and CEO; career summaries from Carnegie Council and IEEE Spectrum also record the Geometric Intelligence acquisition. These roles help explain why his commentary draws on both cognitive science and the practical development of AI systems.
That background is useful context, but it does not make every forecast or policy proposal self-evident. Marcus’s claims are arguments about what current systems can do, what remains unreliable, and what society should require of companies deploying them. Readers can assess those claims separately from his credentials.
#1 Best Overall
- Read Before You Buy — No Video Output: These adapters support charging and USB 2.0 data transfer, but cannot transmit video signals. Except for standard USB webcams (which use USB data only), they are not compatible with HDMI/DisplayPort cables, video-capable USB-C hubs, or docking stations with video output.
- Convert USB-A Ports to USB-C: Designed to connect USB-C earphones, cables, flash drives, card readers, and other USB-C accessories to standard USB-A ports. Plug-and-play with no drivers or software required.
- Aluminum Alloy Housing: Built with a sturdy aluminum alloy shell that aids in heat dissipation and protects against daily wear and scratches. Designed to maintain a stable and secure connection.
- Compact & Travel-Friendly: The ultra-compact design allows the adapter to stay plugged into your device without blocking adjacent ports or adding bulk, reducing wear and tear on your original USB ports.
- 12-Month Warranty: Backed by a 12-month manufacturer warranty for peace of mind. Designed to meet strict quality control standards for reliable everyday performance.
What does Gary Marcus think about ChatGPT and large language models?
Marcus’s central criticism is that a system can be convincing in language while still lacking dependable grounding, common sense, and reasoning. He does not argue that generative AI is useless; he argues against treating fluent output or improvement on selected tasks as proof that a system understands the world or is close to artificial general intelligence.
In an IEEE Spectrum interview, Marcus says generative AI can be useful for bounded purposes such as autocomplete, while warning that it is overvalued when presented as a straightforward route to AGI. His distinction is between a tool that performs a limited task well and a broadly capable system that can reliably transfer knowledge, reason through unfamiliar situations, and act safely.
He identifies several recurring weaknesses in large statistical and deep-learning systems:
- Common sense: they may not reliably apply ordinary background knowledge that people use to interpret everyday situations.
- Causal understanding: recognizing patterns or associations is not the same as knowing what caused an event or what would change if circumstances were different.
- Explicit knowledge: systems may not represent facts in a form that can be inspected, consistently applied, or checked against other knowledge.
- Time and space: a model’s command of language does not guarantee robust understanding of how objects and events persist, move, or relate in the physical world.
- Dependable reasoning: plausible answers are not enough when a system must reach the right conclusion consistently, including in unfamiliar cases.
These are connected problems, not five independent benchmarks. A model that does not maintain a dependable representation of a situation can struggle to reason about what happened, what might happen next, or whether a confident-sounding answer fits the real-world context. Marcus’s point is that language fluency can obscure those gaps rather than resolve them.
His view is especially relevant when someone uses a chatbot’s polished response as evidence of factual reliability. Fluency describes how an answer reads; it does not by itself establish that the answer is grounded, causally sound, or appropriately qualified. A system’s usefulness therefore depends on the task, the consequences of error, and whether its outputs can be checked.
Rank #2
- 5-in-1 USB-C Hub: Experience comprehensive connectivity featuring a Power Delivery input, two USB-A 2.0 ports, a USB-A 3.0 port, and an HDMI port. (Note: The USB-C power delivery input port is only for connecting an external wall charger to power your laptop and cannot power peripheral devices.)
- 90W Pass-Through Charging: Achieve optimal charging with 90W pass-through power to your laptop, supported by a total input of 100W, with the hub reserving 10W for operational efficiency. (Note: Wall charger not included.)
- Quick Data Transfers: Accelerate your productivity with rapid data transfers using a high-speed 5Gbps USB 3.0 port and two 480Mbps USB 2.0 ports.
- 4K HDMI Display: Enhance your visual experience with a hub capable of delivering 4K resolution at 30Hz in both mirror and extend modes. Please note that this hub is compatible with MacBook (macOS 12 and newer), Windows 10 and 11, ChromeOS, and laptops equipped with DP Alt Mode and Power Delivery. Note: This device is not compatible with Linux.
- What You Get: Anker USB-C Hub (5-in-1, 4K HDMI), welcome guide, 18-month warranty, and our friendly customer service.
Does Marcus think scaling alone will lead to AGI?
No. Marcus doubts that simply making large learning systems bigger or training them on more data will be enough to produce human-level general intelligence. In the IEEE Spectrum interview, he states, “I don’t think we’re anywhere near AGI.” He also says, “I suspect we need to take a broad view of how the human mind works if we’re to build really advanced AI.”
Those remarks capture an important distinction: Marcus is not saying that future advanced AI is impossible. He is challenging the assumption that current successes demonstrate a reliable path to it. In his view, progress requires taking seriously the range of capacities involved in intelligence, rather than equating scale or impressive language performance with general understanding.
That position is a critique of a scaling-only forecast, not a claim that every large model is ineffective. His acknowledgment of bounded uses such as autocomplete leaves room for practical value even while he disputes ambitious claims about what these systems understand or can become.
What does Marcus propose instead of deep learning alone?
Marcus favors hybrid, sometimes called neurosymbolic, approaches that combine learning from data with structured representations and ways to verify conclusions. The basic idea is to use statistical learning where it is effective while adding mechanisms that can represent explicit knowledge, support reasoning, and make results more checkable.
This is a direction rather than a single recipe. The cited profiles and interviews do not establish one architecture or implementation as Marcus’s complete solution to AI’s limitations. His broader argument is that systems need more than pattern learning if they are to handle common sense, causality, time, space, and reliable reasoning.
Rank #3
- Sleek 7-in-1 USB-C Hub: Features an HDMI port, two USB-A 3.0 ports, and a USB-C data port, each providing 5Gbps transfer speeds. It also includes a USB-C PD input port for charging up to 100W and dual SD and TF card slots, all in a compact design.
- Flawless 4K@60Hz Video with HDMI: Delivers exceptional clarity and smoothness with its 4K@60Hz HDMI port, making it ideal for high-definition presentations and entertainment. (Note: Only the HDMI port supports video projection; the USB-C port is for data transfer only.)
- Double Up on Efficiency: The two USB-A 3.0 ports and a USB-C port support a fast 5Gbps data rate, significantly boosting your transfer speeds and improving productivity.
- Fast and Reliable 85W Charging: Offers high-capacity, speedy charging for laptops up to 85W, so you spend less time tethered to an outlet and more time being productive.
- What You Get: Anker USB-C Hub (7-in-1), welcome guide, 18-month warranty, and our friendly customer service.
For people evaluating an AI product, this argument suggests practical questions beyond “How impressive is its output?” Ask whether it can explain or expose relevant information in a form that can be checked, whether it is evaluated on the actual task, and what happens when it is wrong. Those questions do not guarantee safety, but they focus attention on dependability rather than presentation.
Do these 3 things before closing this tab:
1Scan for outdated or missing drivers - takes under a minute2Clear out junk files and repair common Windows errors3Fix the driver behind crashes, sound loss and screen glitchesWhy does Marcus call for AI oversight?
Marcus’s concerns extend beyond technical capability. He warns that AI can intensify social and political harms, particularly when powerful systems are deployed without transparency, accountability, or meaningful public scrutiny. His policy argument is that beneficial applications should remain possible, but companies should not be the sole arbiters of acceptable risk.
In the IEEE Spectrum interview, he raises concerns about misinformation and deepfakes, including deepfake pornography, as well as copyright disputes over training data, opaque system behavior, weak accountability, and concentrated corporate power. He also worries that technology companies can shape public policy in ways that serve their own interests. These concerns connect model behavior to the institutions that build, deploy, and govern the models.
His 2024 book, Taming Silicon Valley: How We Can Ensure That AI Works for Us, sets out his policy-focused case. MIT Press describes the book as addressing Big Tech influence and AI’s potential harms, with policy tools including data rights, layered AI oversight, transparency, liability, and tax reform. The emphasis is not on banning every application; it is on creating rules and accountability that can constrain harms while leaving room for useful AI.
“Layered” oversight matters because AI systems can create different kinds of risk in different settings. A low-stakes writing aid and a system used in a consequential decision do not call for identical scrutiny. Marcus’s policy framework, as summarized by the publisher, includes several levers rather than relying on a single fix: rights around data, transparency about systems, liability for harms, and oversight that can adapt to the deployment.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Rank #4
- Dual Converters, Infinite Potential：Includes 2× USB C male to USB A female adapters and 2× USB A male to USB C female adapters. Perfect for a wide range of uses—tablets with Bluetooth keyboards, expand USB ports on macbook, and more. Two different converters for all your daily needs
- Next-Level 10Gbps & 3A Charging: No more slow 480Mbps, this usb to usb c adapter has a transfer speed of up to 10Gbps, allowing you to do more transferring in less time. This usb adapter fits both USB A and USB C charger, supporting up to 3A fast charging
- Upgraded Exquisite Craftsmanship: With an aluminum alloy housing and metal connector, the usbc to usb adapter is extremely durable and sturdy. Rigorously tested to withstand more than 10,000 times of plugging and unplugging, ensuring long-lasting performance
- Broad Compatible: The usb c to usb adapter widely supports all USB C/ USB A devices like laptops, tablets, cellphones, car chargers, and phone chargers. Such as compatible with MacBook Pro/Air 2023/2022, Thunderbolt 4/3 Devices,Apple MagSafe Watch 9/8/7/SE/Ultra, iPad Pro 2022/2021, Samsung Galaxy S23/S20/S10, and iPhone 17/16/15 Pro. Plug and play
- Please Note: To reach 10Gbps speed, keep the cable under 3.3 ft. For USB A Male to USB C adapters, try flipping the USB C connector. USB C Male to USB A adapters support bidirectional 10Gbps transfer within 3.3 ft
What should you read first by Gary Marcus?
The best starting point depends on whether you want a general explanation of AI’s technical limits, a policy argument, or his latest commentary. The sequence below moves from foundational diagnosis to governance and then to ongoing debate.
- Rebooting AI: Building Artificial Intelligence We Can Trust, with Ernest Davis. Start here for a general-audience account of common sense, causal models, and trustworthy-system design. It is the clearest entry point in the recommended reading path for understanding Marcus’s technical critique.
- Taming Silicon Valley: How We Can Ensure That AI Works for Us. Read this next for his policy and governance proposals, including data rights, oversight, transparency, liability, and democratic accountability. MIT Press gives its publication date as September 17, 2024, and lists it at 248 pages.
- “Deep Learning: A Critical Appraisal” and “Deep Learning Is Hitting a Wall.” These offer, respectively, a technical paper and a general-audience treatment of his critique of approaches that rely on scaling deep learning alone. They are useful after the books if you want to follow the argument in more technical or essay-length form.
- Marcus on AI. For reactions to new model releases, AGI claims, AI economics, safety, and policy developments, use his Substack publication and its archive. The homepage is for recent posts; the archive lets readers locate dated essays.
This order is not a claim that one book settles the debate. It is a way to separate the questions Marcus addresses: what current systems struggle with, what technical approaches might help, and what rules should govern deployment.
How do Marcus’s books, papers, and newsletter differ?
His work varies by audience and purpose. A book offers a broader argument, a paper focuses on a technical critique, and newsletter posts respond to a changing field. The following comparison is based on the works and descriptions identified by their publishers and institutional profiles.
| Work or outlet | Main focus | Best suited to | Details established by the cited sources |
|---|---|---|---|
| Rebooting AI (with Ernest Davis) | Common sense, causality, and trustworthy AI design | Readers seeking a broad, accessible account of technical limitations | Exact publication date, page count, and price are not stated in the cited profiles. |
| Taming Silicon Valley | AI harms, technology-company power, and policy tools | Readers interested in governance and democratic accountability | MIT Press lists September 17, 2024; 248 pages; paperback ISBN 9780262551069; eBook ISBN 9780262381567; and a $18.95 paperback price in its 2024 listing. |
| “Deep Learning: A Critical Appraisal” | Technical assessment of deep learning’s limits | Readers who want a technical treatment | It is identified as a technical paper; publication and access details are not stated in the cited source. |
| “Deep Learning Is Hitting a Wall” | General-audience critique of scaling-centered approaches | Readers who want an essay-length introduction to the critique | Publication and access details are not stated in the cited source. |
| Marcus on AI | Commentary on LLM reliability, AGI claims, economics, safety, and policy | Readers following current debate and new developments | Substack publication with a homepage and archive; specific subscription terms are not stated in the cited sources. |
MIT Press’s $18.95 figure is the publisher’s listed paperback price in its 2024 information, not a guarantee of the price at every retailer or at a later date. The publisher identifies the book’s paperback and eBook ISBNs separately, which can help when checking format or locating a specific edition.
Windows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallCrashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minuteWhere does Gary Marcus publish and where can you find background?
Marcus’s Substack, Marcus on AI, is the most direct place to follow his ongoing commentary. Its posts address fast-moving subjects such as LLM reliability and new AGI claims, while the dated archive is more useful when researching a particular argument.
Best Value
- 5-in-1 Connectivity: Equipped with a 4K HDMI port, a 5 Gbps USB-C data port, two 5 Gbps USB-A ports, and a USB C 100W PD-IN port. Note: The USB C 100W PD-IN port supports only charging and does not support data transfer devices such as headphones or speakers.
- Powerful Pass-Through Charging: Supports up to 85W pass-through charging so you can power up your laptop while you use the hub. Note: Pass-through charging requires a charger (not included). Note: To achieve full power for iPad, we recommend using a 45W wall charger.
- Transfer Files in Seconds: Move files to and from your laptop at speeds of up to 5 Gbps via the USB-C and USB-A data ports. Note: The USB C 5Gbps Data port does not support video output.
- HD Display: Connect to the HDMI port to stream or mirror content to an external monitor in resolutions of up to 4K@30Hz. Note: The USB-C ports do not support video output.
- What You Get: Anker 332 USB-C Hub (5-in-1), welcome guide, our worry-free 18-month warranty, and friendly customer service.
For book details, MIT Press is the authoritative source for Taming Silicon Valley’s publication information, formats, and ISBNs. IEEE Spectrum has an author page and an accessible interview covering his biography and positions on generative AI. Carnegie Council and institutional biographies provide additional background on his career and academic work.
For earlier and more general-audience writing, institutional profiles identify his contributions to The New Yorker and The New York Times. Frontier Minds also points readers to “Deep Learning Is Hitting a Wall,” “Deep Learning: A Critical Appraisal,” his 2023 Senate testimony, and Rebooting AI. These sources serve different purposes: current commentary, book information, accessible interviews, or a longer treatment of a technical argument.
How should readers assess Marcus’s arguments?
Read his work as a sustained challenge to the inference that fluent output equals robust intelligence, and as a call to match AI claims and safeguards to the system’s actual use. His technical and policy writing are related, but they answer different questions: one asks what systems can reliably do; the other asks who should bear responsibility and what oversight should exist.
The Tool Desk
Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →- Separate task performance from general capability. A system may be useful for a bounded task without demonstrating broad, reliable understanding.
- Ask what counts as evidence. Fluency or a striking example does not settle whether a model handles unfamiliar cases, causal questions, or real-world context consistently.
- Distinguish technical proposals from policy tools. Hybrid architectures address how systems might be built; transparency, liability, data rights, and oversight address how systems are governed.
- Keep the stakes in view. The consequences of an error and the ability to verify an output matter when deciding how much trust to place in a system.
Marcus’s skepticism about scaling-only paths to AGI should not be mistaken for a claim that AI has no present value. Nor does his concern about harms imply that every application should be prohibited. His position, as represented in his books and interviews, is that practical usefulness, technical limits, and social accountability have to be considered together.
Frequently Asked Questions
Did Gary Marcus testify to the U.S. Senate about AI?
Yes. Frontier Minds includes his 2023 Senate testimony among the materials it recommends for readers exploring his policy views. It is a useful primary-source option if you want to see how he presented those concerns in a government hearing.
Has Gary Marcus written for The New Yorker or The New York Times?
Yes. Institutional biographies identify contributions to both publications. Those outlets are part of his earlier general-audience writing, distinct from his current commentary feed.
What did Gary Marcus study before becoming an AI commentator?
His studies included cognitive science at Hampshire College and MIT, and he later taught psychology and neural science at NYU. His path helps explain why his AI writing often draws comparisons between machine learning and human cognition.
Recommended Free Tools
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
