Soon after the first proofs of concept for AI agents, Simon Willison shares his concerns about how people are hooking LLMs up to external tools

Dell'Acqua Fabrizio, Kogut Bruce and Perkowski Patryk demonstrate experimentally that even in a task where AI outperforms humans, the replacement of a human player by an automated videogame agent decreases team performance.

Jonathan T. Barron, Ben Mildenhall, Dor Verbin, Pratul P. Srinivasan and Peter Hedman show how ideas from rendering and signal processing can be used to construct a technique that combines mip-NeRF 360 and grid-based models such as Instant NGP to yield error rates that are 8%-77% lower than either prior technique, and tha...

Wired - Tim Marshall explores China’s growing influence in space through its sci-fi narratives, technological advancements and political agenda, as detailed in his book 'The Future of Geography'

Big Think - With a massive, charged nucleus orbited by tiny electrons, atoms are such simple objects. Miraculously, they make up everything we know.

Ed Felten, Manav Raj and Robert Seamans identify and characterize the occupations and populations most exposed to advances in generative AI, and documents how policy interventions may be necessary to help individuals adapt to the new demands of the labor market.

The IDL documentation provides insightful details on an efficient way to represent points in a 3D-space.

Joon Sung Park, Joseph C. O'Brien, Carrie J. Cai, Meredith Ringel Morris, Percy Liang and Michael S. Bernstein instantiate generative agents to populate an interactive sandbox environment inspired by The Sims, where end users can interact with a small town of twenty five agents using natural language.

Alexander Pan, Jun Shern Chan, Andy Zou, Nathaniel Li, Steven Basart, Thomas Woodside, Jonathan Ng, Hanlin Zhang, Scott Emmons and Dan Hendrycks introduce MACHIAVELLI, a benchmark of 134 Choose-Your-Own-Adventure games containing over half a million rich, diverse scenarios that center on social decision-making.

Alexander Kirillov, Alex Berg, Chloe Rolland, Eric Mintun, Hanzi Mao, Laura Gustafson, Nikhila Ravi, Piotr Dollar, Ross Girshick, Spencer Whitehead and Wan-Yen Lo introduce the Segment Anything (SA) project: a new task, model, and dataset for image segmentation.

WIRED - Digging through manuals for security cameras, a group of gearheads found sinister details and ignited a new battle in the US-China tech war.

BOOK | This book by Maxime Labonne, covers the fundamental concepts of graph neural networks and their implementation using Python and PyTorch Geometric. It includes techniques for classifying nodes, graphs, and edges, as well as predicting and generating realistic graph topologies. The book also discusses real-world applicat...

Jia Nan, Luo Xueming, Fang Zheng and Liao Chengcheng analyze the impact of AI assistants on employee creativity

Volunteers from Ontocord.AI, Bedrock AI, TurkuNLP, ETH, Redmond.AI, Incite, MICS CentraleSupelec, Centro de Excelência em Inteligência Artificial, VietAI, Technion - Israel Institute of Technology, Nous Research, University of Western Australia, LAION.AI, KoboldAI Community, MILA, Luleå University of Technology, Juelich Supe...

Yongliang Shen, Kaitao Song, Xu Tan, Dongsheng Li, Weiming Lu and Yueting Zhuang se ChatGPT to conduct task planning when receiving a user request, select models according to their function descriptions available in Hugging Face, execute each subtask with the selected AI model, and summarize the response according to the exec...

Quantamagazine - David S. Richeson delves into the history of the four-color problem, which asks whether it is always possible to color any map using at most four colors while ensuring that regions sharing a border are different colors.

Tyna Eloundou , Sam Manning, Pamela Mishkin and Daniel Rock investigate the potential implications of large language models (LLMs), such as Generative Pretrained Transformers (GPTs), on the U.S. labor market, focusing on the increased capabilities arising from LLM-powered software compared to LLMs on their own

Fabrizio Gilardi, Meysam Alizadeh, and Maël Kubli found that ChatGPT outperformed crowd-workers in text annotation tasks such as relevance, stance, topics, and frames detection. The model's zero-shot accuracy was higher for four out of five tasks, and its intercoder agreement exceeded both crowd-workers and trained annotators...

Suchin Gururangan, Margaret Li, Mike Lewis, Weijia Shi, Tim Althoff, Noah A. Smith and Luke Zettlemoyer introduce a simple but effective method to asynchronously train large, sparse language models on arbitrary text corpora.

OpenAi's GPT4 System Card shows that the company did look into risks and checked that the most imminent threats were appropriately addressed.

Sébastien Bubeck, Varun Chandrasekaran, Ronen Eldan, Johannes Gehrke, Eric Horvitz, Ece Kamar, Peter Lee, Yin Tat Lee, Yuanzhi Li, Scott Lundberg, Harsha Nori, Hamid Palangi, Marco Tulio Ribeiro and Yi Zhang discuss the rising capabilities and implications of the latest models and demonstrate that, beyond its mastery of langu...

Authors propose a novel model architecture, Receptance Weighted Key Value (RWKV), that combines the efficient parallelizable training of Transformers with the efficient inference of RNNs.

Bill Gates announces the beginning of a new era, explaining that Artificial intelligence is as revolutionary as mobile phones and the Internet.

J. Curtis Nickel and R. Christopher Doiron present a novel sublingual mucosal-based bacterial vaccine developed for the prevention of recurrent UTI in women

LanceDB is an open-source vector database designed for AI applications. Built on a scalable columnar data format called Lance, it supports large-scale multi-modal data storage, management, querying, and retrieval of embeddings.