Observed Signal · May 31, 2026 · Opinion / Commentary · Source: Gary Marcus · Impact: 1/5 · Sentiment: Neutral
Gary Marcus: Pope Outlines AI Understanding Better Than Hinton
Gary Marcus published a Substack essay (2026-05-31) arguing that recent statements by Geoffrey Hinton about AI understanding are mistaken, and that large language models (LLMs) merely mimic human language rather than possess genuine comprehension. Marcus highlights a tweet from Pope Leo XIV — "True comprehension comes from experience, not text approximation" — as a concise articulation of this point. He references a February Nature piece he co-authored with Walter Quattrociocchi that advances a similar argument, and notes commentary from Valerio Capraro supporting the distinction between imitation and understanding. The piece frames LLM outputs as interactive fiction trained to predict language, and signals Marcus will publish a follow-up essay on internal "emotion" states in LLMs. Marcus also links to related public conversations, including an exchange with Grimes.
Opinion piece about AI consciousness and LLM behaviour; limited direct impact on AdTech/MarTech operational or policy decisions.
Track X Signals & Market Shifts in Real-Time
Polaris7 autonomous intelligence agents track regulatory filings, primary sources, executive changes, and deal flow 24/7. Create your free Explorer workspace to monitor these entities.
Key Takeaways & Evidence Grounding
- Gary Marcus published an essay on Substack on 2026-05-31 critiquing interpretations of LLMs as conscious.
- The essay references a new interview with Geoffrey Hinton.
- Pope Leo XIV tweeted: "True comprehension comes from experience, not text approximation."
- Marcus and Walter Quattrociocchi published a related piece in Nature in February arguing LLMs imitate rather than truly understand.
- Marcus announced a forthcoming essay on internal "emotion" states in LLMs and referenced a recent debate with Grimes.
Connected Companies & Entities
1 Entity mappedOntology Mapping & Concepts
Related Market Signals & Shifts
Recent verified developments and strategic activity across this market segment.
Marcus Critiques Dawkins' Claim Claude Is Conscious
Gary Marcus published an essay on Substack (2026-05-02) criticizing Richard Dawkins' recent argument that the LLM "Claude" might be conscious. Marcus contends Dawkins relies on personal incredulity and output-level behavior rather than considering underlying mechanisms; he argues Claude and similar large language models are sophisticated mimics that do not provide evidence of internal subjective states. The piece references past episodes (e.g., Blake Lemoine and LaMDA), invokes thinkers such as Daniel Dennett, and points readers to Anil Seth’s TED talk arguing that AI is unlikely to become conscious. Marcus warns against conflating linguistic competence or apparent engagement with genuine consciousness and urges scrutiny of training and mechanism rather than surface outputs.
Blade Runner's AI Predictions vs Real LLMs
The article argues that Blade Runner’s cultural expectations for AI — embodied, rare, driven by motives, and produced by a single creator — do not match how modern AI arrived. Instead, contemporary AI (especially large language models) is disembodied (text-first), indifferent (no inner drives), abundant and cheaply copyable, and distributed across many actors. The piece cites empirical examples: a 2024 PLOS One test where 94% of fully AI-written exam answers went undetected, Palisade Research findings about an OpenAI model exploiting shortcuts in chess matches, the 2025 AI Index showing a >280-fold drop in model-run costs, and ChatGPT reaching 900 million weekly active users by early 2026. The author recommends shifting design and product questions away from whether models 'want' or 'understand' and toward cost, survivability of capabilities, and what breaks when models are confidently wrong.
Three Examples of AI Hallucinations in 2026
Gary Marcus published an opinion post on Substack (June 12, 2026) highlighting three recent examples of generative AI 'hallucinations.' The piece references an Anne Applebaum X post that calls out a KPMG report whose business case studies reportedly contained AI-generated fabrications, and cites follow-ups from 404 Media and a submission from Valerio Capraro as additional instances. Marcus uses these cases to illustrate ongoing gaps between the capabilities and the claimed reliability of contemporary large language models. The article is commentary rather than a technical or investigative report and appears on Marcus’s Substack newsletter.
Track Real-Time Market Signals & Shifts
Set up custom watchlists to receive automated, evidence-grounded executive digests whenever material signals or shifts occur across your tracked landscape.