OpenAI CEO Sam Altman expresses concern about the use of AI to interfere with election integrity and calls for regulation and guidelines to be put in place.
Altman suggests that the U.S. consider the implementation of licensing and testing requirements for development of AI models and believes a model that can manipulate a person's beliefs should be subject to a "great threshold" for licensing.
Altman also believes that companies should have the right to refuse the use of their data for AI training, but material on the public web should be fair game.
There is ongoing debate on AI regulation and safety, with some suggesting the potential for extreme outcomes and others proposing approaches such as public registration requirements for training data and model.
The primary concern is the need for a democratic and transparent approach to regulation, with focus on ensuring safety once AIs are widely used, and considering potential impacts on public infrastructure and employment.
JPMorgan Chase & Co's employee surveillance tool called WADU has raised concerns among employees and potential hires. The AI & machine learning system uses facial and speech recognition to track employees and monitor non-verbal body language in real-time.
Upper management prefers employees to work in the office to refine the WADU profiles, even if employees may be more productive working from home.
JPMorgan Chase's WADU program collects a vast amount of data on employees, including monitoring keystrokes, mouse clicks, cameras, and microphones, alerting managers if specific content is detected. Concerns have been raised about the program's extreme surveillance tactics.
Critics argue that such measures represent an invasion of privacy and raise concerns about employee productivity and well-being, while others suggest these technologies are common in the banking industry and mandated by regulators.
The post sparks discussion about workplace surveillance, the use of metrics to measure productivity, and the pressure to constantly work, with concerns about privacy breaches and negative impacts on employee well-being.
Exposure to the chemical solvent trichloroethylene (TCE) increases the risk of Parkinson's disease, according to an epidemiological study published in JAMA Neurology.
Veterans who trained at Marine Corps Base Camp Lejeune in North Carolina from 1975 to 1985 and were exposed to water with high concentrations of TCE had a 70% higher risk of developing Parkinson's disease compared to those who trained elsewhere.
TCE is a colorless liquid used mainly to produce refrigerants and as a degreaser in heavy industry. It is highly persistent in soil and groundwater, detectable in food, and present in up to one-third of US drinking water.
The chemical was once common in various applications, including decaffeinated coffee and as an anesthetic for children and women in labor, but its use is now highly restricted.
TCE persists in soil and groundwater, and while it is not detectable in most retail products, it is detectable in up to one-third of US drinking water.
Apple previews new software features for users with cognitive, vision, hearing, and mobility accessibility.
Assistive Access, Live Speech, and Personal Voice will be released later this year to increase ease and independence for users with cognitive disabilities.
Point and Speak in Magnifier, a feature that helps users with vision disabilities interact with physical objects, has been introduced.
Apple's plan to move LLM models like GPT to the edge and offer complete privacy with its silicone architecture puts it in a good position to compete with other big tech companies' cloud-based AI offerings and sets it apart in users' eyes as a champion of privacy.
Apple has announced new accessibility features, including Personal Voice, Assistive Access, eye tracking, and limited mobility features, that make everyday technology accessible and user-friendly for everyone and shows that narratives matter as a counterweight to the fear surrounding AI and automation.
Guidance is a new program for controlling Language Model Controllers, with efficient language model control, API-based chat models, and smart seed-based generation caching.
Guidance offers rich output structures, including multiple generations, selections, conditionals, and tool-use. It also has a guarantee of valid syntax with JSON, role-based chat model examples, and a unified API based on role tags.
The concern of prompt injection attacks is raised.
Microsoft has also released a new Python-based framework called Guidance, which facilitates the use of large language models (LLMs) and provides developers leverage over other tools in their machine learning toolkit.
The Sourdough Framework is an open-source book that provides a framework for making the best sourdough bread at home
The book gives information on the natural fermentation process with scientific references to help bakers in different environments
The book can be built locally or downloaded in PDF, EPUB, MOBI, or AZW3 versions, and readers can contribute to improving the book via Github or donate to support the author.
The project called "Open Sourdough" aims to make baking bread accessible by documenting the process with Git and Docker, but some sections might need more clarity.
Several tips for successful sourdough baking were shared, such as using a fridge for storage, measuring ingredients by weight, and using a Dutch oven.