Knowledge catalogue
arXiv:2605.22263v1 Announce Type: new Abstract: On-policy self-distillation (OPSD) is an emerging LLM post-training paradigm in which the model serves as its own teacher: conditioned on privileged inf
arXiv:2605.22376v1 Announce Type: new Abstract: Cross-domain offline reinforcement learning (CDRL) aims to improve policy learning in a target domain by leveraging data collected from a source domain.
arXiv:2605.21492v1 Announce Type: new Abstract: No feature ranking can be simultaneously faithful, stable, and complete when features are collinear. For collinear pairs, ranking reduces to a coin flip
The biggest myth of all is that I am the only one who doubts the hype. Demonize me all you like, but the bubble is still probably gonna burst. So many other people see it, too. Speculative hype bubble
Gary Marcus comments on OpenAI's financial losses, referencing a period when the company's annual losses were approximately $5 billion as a point of comparison to its current financial situation. The
arXiv:2605.22800v1 Announce Type: new Abstract: Robustness, domain adaptation, photometric and occlusion invariance, compositional generalisation, temporal robustness, alignment safety, and classical
the OpenAI crowd is suddenly coming after me relentlessly, but too chicken to actual face me in a moderated debate. here’s why: a. the IPO is coming, but OpenAI’s has lost their lead over Anthropic an
arXiv:2605.22496v1 Announce Type: new Abstract: Deep generative models offer a natural foundation for out-of-distribution (OOD) detection, yet prior work has shown that their assigned likelihoods are
The US government absolutely should NOT bail out OpenAI. They have AFAIK been given more funding than anyone in history and there is no evidence they can run a profitable business,and meanwhile many o
This is collectively the largest financial iceberg ever put in front of the stock market. This is $5 trillion in TOTAL AIR. These companies have never made a dime. SpaceX is literally a government wel
🦔This one happened yesterday but is still worth flagging. Starbucks killed its AI-powered inventory counting tool after nine months in North American stores. The system used LiDAR sensors and cameras
Three cheers for @flowersslop who agrees with me about nothing but still sees the value of thoughtful disagreement. thanks feels like a hot take now, but any space needs a real variety of opinions. I
Total amount raised: 190 billion Total spending pledged: 600 billion Total profits (cumulative): 0 @CarinaN818 @GaryMarcus Because you can raise 190 billion in funding by releasing an LLM chatbot that
arXiv:2306.05905v2 Announce Type: replace Abstract: A convenient approach to optimally solving combinatorial optimization tasks is the Branch-and-Bound method. Its branching heuristic can be learned t
arXiv:2511.14220v3 Announce Type: replace Abstract: Model-based reinforcement learning (RL) methods that leverage search are responsible for many milestone breakthroughs in RL. Sequential Monte Carlo
arXiv:2502.13822v3 Announce Type: replace-cross Abstract: We establish novel and general high-dimensional concentration inequalities and Berry-Esseen bounds for vector-valued martingales induced by Ma
Gary Marcus discusses the risks that AI systems may function adequately until they unexpectedly fail catastrophically, potentially causing severe economic damage. The post suggests that current AI rel
Gary Marcus reports on X that someone named Theo has clarified he did not accept fees from OpenAI, following earlier ambiguity about his financial relationships with the company. The post appears to b
arXiv:2605.21507v1 Announce Type: cross Abstract: Atmospheric visibility is a critical variable for transportation safety and air quality management, however, accurate prediction remains challenging d
Gary Marcus critiques OpenAI's business model sustainability, suggesting it lacks transparency or clear long-term viability beyond investor confidence. The post implies skepticism about OpenAI's path
arXiv:2605.22257v1 Announce Type: new Abstract: Formal theorem provers based on large language models (LLMs) are highly sensitive to superficial variations in problem representation: semantically equi
arXiv:2605.21868v1 Announce Type: new Abstract: In competitive games, players frequently switch strategies after losing streaks, yet our analysis of 926,334 match records from 34,619 Clash Royale play
World models have existed for years (though not in LLMs); I take them to be explicit representation of objects, places, events, mechanisms etc you can reason over. Chess computers have them (board, pi
you can destroy a person’s simply by lying about them. that is what OpenAI is trying do to me. ask yourself why The case against me below is completely intellectually dishonest, filled with lies and m
100% agree with @BethMayBarnes on this point and most (though not quite all*) of her important thread. *i am much less concerned about extinction risk per se, as discussed in my TLS review of If Anyon
arXiv:2510.04280v2 Announce Type: replace-cross Abstract: Effective exploration remains a central challenge in model-based reinforcement learning (MBRL), particularly in high-dimensional continuous co
arXiv:2605.22608v1 Announce Type: new Abstract: Agentic systems are becoming more capable: agents define strategies, take actions, and interact with different environments. This autonomy poses serious
Gary Marcus announced an upcoming CNBC live segment scheduled for approximately 11:15 ET featuring hosts Carl Quintanilla and Leslie Picker to discuss the intersection of artificial intelligence and i
America’s biggest companies have gone from printing money to burning it. It does not take Poirot to work out what’s going on. Register for free to get our columnist’s take on it http://econ.st/4tKbpoB
arXiv:2605.21653v1 Announce Type: cross Abstract: AI text detectors amplify a pretrained typicality axis; they do not construct an AI-vs-human boundary. On raw encoders before any task supervision, pr
arXiv:2605.22123v1 Announce Type: new Abstract: Designing reward functions that generalize beyond controlled laboratory settings remains a fundamental challenge in reinforcement learning for robotics.
arXiv:2605.21957v1 Announce Type: new Abstract: Video anomaly detection is critical for public safety and security, yet remains highly challenging despite extensive research due to large variations in
arXiv:2605.22600v1 Announce Type: new Abstract: Motion planning for autonomous driving must account for multi-modal uncertainty in both the intentions and trajectories of surrounding vehicles. Handlin
arXiv:2605.22547v1 Announce Type: new Abstract: Deep learning has brought significant progress to medical image classification, yet most existing methods still rely on isolated visual evidence and can
arXiv:2505.16416v3 Announce Type: replace Abstract: Rotary Position Embedding (RoPE) is widely adopted in large language models, but when applied to vision-language models (VLMs) it couples text and i
arXiv:2605.21688v1 Announce Type: new Abstract: Autonomous contact-based micromanipulation is challenging because surface and interfacial interactions at the microscale are difficult to model accurate
arXiv:2605.21609v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly embedded in adolescent digital environments, mediating information seeking, advice, and emotionally sensit
arXiv:2602.17517v2 Announce Type: replace Abstract: Augmented reality (AR) guidance in laparoscopic liver surgery requires accurate registration of preoperative 3D models to intraoperative 2D video, b
arXiv:2605.21958v1 Announce Type: new Abstract: When a multi-module LLM agent fails, the module most responsible for the failure is not necessarily the best place to intervene. We demonstrate this Dia
arXiv:2602.15338v2 Announce Type: replace-cross Abstract: Large language model (LLM) alignment relies on complex reward signals that often obscure the specific behaviors being incentivized, creating c
arXiv:2605.22061v1 Announce Type: new Abstract: Distributed Image Compression (DIC) is crucial for multi-view transmission, especially when operating at extremely low bitrates (< 0.1 bpp). Its core ch
Gary Marcus comments positively on what appears to be a strong period for initial public offerings (IPOs) in a brief social media post. The post suggests optimism about market conditions favorable to
arXiv:2605.22066v1 Announce Type: new Abstract: Reconstructing 4D (3D+t) cardiac geometry from sparse 2D echocardiography is highly desirable yet fundamentally challenged by geometric ambiguity and te
Eight key points from the most recent essay in the “AI as Normal Technology” series by @sayashk and me. Do AI Risks Require Extraordinary Government Intervention? 1. There is general consensus that AI
arXiv:2605.21842v1 Announce Type: cross Abstract: Standard transformer attention computes pairwise similarity between queries and keys, treating all tokens as equally salient regardless of their intri
arXiv:2605.22185v1 Announce Type: new Abstract: Recent advancements in Multimodal Large Language Models (MLLMs) have demonstrated impressive capabilities in general visual understanding. However, thei
arXiv:2605.22057v1 Announce Type: new Abstract: Enterprise routers assign queries to expert agents, yet deployed profiles stay static while agents evolve (prompts, tools, models), and developers rarel
arXiv:2602.17186v2 Announce Type: replace Abstract: Large Vision Language Models (LVLMs) have achieved remarkable progress, yet they often suffer from language bias, producing answers without relying
arXiv:2605.22671v1 Announce Type: new Abstract: Vision-Language-Action (VLA) models often suffer from performance degradation under distribution shifts, as they struggle to learn generalized behavior
Fuck this OpenAI employee, seriously fuck him. Also read this quantitative study, which I had nothing to do with, that says my technical predictions have bee largely correct. https://github.com/davego
“Gary Marcus, former CEO of Geometric Intelligence and New York University professor, joins @SquawkStreet @CNBC to discuss why he still believes OpenAI could be the ‘WeWork’ of AI, why he finds Anthro
arXiv:2605.21605v1 Announce Type: new Abstract: Open-ended image generation is no longer a simple prompt-to-image problem. High-quality generation often requires an agent to combine a model's internal
arXiv:2605.22619v1 Announce Type: new Abstract: Grounding radiology report descriptions to 3D CT volumes is essential for verifiable clinical interpretation, yet remains challenging due to the semanti
Lauren Feiner / The Verge: Google files its appeal of the US federal ruling deeming it an illegal search monopolist, arguing it “prevailed in the marketplace fair and square” — It wants to throw out t
arXiv:2605.20874v1 Announce Type: new Abstract: Enterprise agents are increasingly expected to operate autonomously across tools and interfaces, yet production deployments require governance by constr
arXiv:2605.20210v1 Announce Type: cross Abstract: Agentic AI systems - systems that can pursue goals through multi-step planning and tool-mediated action with limited direct supervision - are moving f
arXiv:2605.22258v1 Announce Type: new Abstract: Large language models (LLMs) require robust toxicity evaluation beyond explicit wording. This setting remains underexplored in Chinese, where toxicity m
arXiv:2605.21661v1 Announce Type: cross Abstract: Adapting pretrained diffusion models to downstream objectives such as inverse problems often requires expensive test-time guidance or optimization. We
arXiv:2605.22322v1 Announce Type: new Abstract: Reasoning capability has significantly advanced complex logical inference and robotic decision-making in general domains. However, its potential in the
I just ran into a guy I know in NYC who said he’s doing a temporary gig for OpenAI. According to him, OpenAI is paying hundreds of people and families in New York to install 360-degree cameras through