The Cluster Project is an art group engaged in making an ongoing series of multimedia works that explore the thriving universe of drones, nukes, cluster bombs, and other indiscriminate weapons of terror. These videos, animations, illustrations, performances, interventions, and dataflow programming works strive to illuminate our culture’s enduring embrace of — and simultaneous disinterest with — weapons and war.
Source: About – The Cluster Project
We focused the RSP on the principle of conditional, or if-then, commitments. If a model exceeded certain capability levels (for example, biological science capabilities that could assist in the creation of dangerous weapons), then the policy stated that we should introduce a new and stricter set of safeguards (for example, against model misuse and the theft of model weights).
Source: Responsible Scaling Policy Version 3.0 Anthropic
The U.S. Navy has a large collection of marine mammals that are used as military working dogs of the ocean. The Navy also studies the diving physiology, hydrodynamics, and sonar capabilities of marine mammals. Whale and dolphin sonar is studied because nothing can be developed electronically that could mimic these living systems.
Source: Read “Scientific Advances in Animal Nutrition: Promise for the New Century: Proceedings of a Symposium” at NAP.edu
aTrain is a tool for automatically transcribing speech recordings utilizing state-of-the-art machine learning models without uploading any data. It was developed by researchers at the Center for Data Science in Business and Society – BANDAS at the University of Graz and tested by researchers from the Know-Center Graz.
Source: aTrain – Center for Data Science in Business and Society – BANDAS
Semantic ablation is the algorithmic erosion of high-entropy information. Technically, it is not a “bug” but a structural byproduct of greedy decoding and RLHF (reinforcement learning from human feedback).
Source: Semantic ablation: Why AI writing is boring and dangerous • The Register
For most of the history of software engineering, we’ve built systems around a simple and comforting assumption: Given the same input, a program will produce the same output. When something went wrong, it was usually because of a bug, a misconfiguration, or a dependency that wasn’t behaving as advertised. Our tools, testing strategies, and even our mental models evolved around that expectation of determinism.AI quietly breaks that assumption.
Source: AI Is Not a Library: Designing for Nondeterministic Dependencies – O’Reilly
The Transcription powered by Whisper.cpp can both transcribe and translate words and outputs to a label track. If you want to export these transcriptions, you can do so via File ? Export Other ? Export Labels.
Source: Audacity ® | Introducing OpenVINO AI effects for Audacity
h/t Joe A.