Thursday☕️

Earth Intelligence

30 Nov 2023 — 4 min read

Trending:

France Smoking Rates
AWS AI Testing

Markets:

Yesterday’s U.S. stock market:

Yesterday’s commodity market:

Yesterday’s crypto market:

France Smoking Rates:

France is taking significant steps towards establishing a "tobacco-free generation" by 2032. In a country where the average adult smokes 1,089 cigarettes annually, and approximately 15% of deaths each year are attributed to smoking, these measures represent a major public health initiative.
The French health minister recently unveiled a comprehensive plan to curb smoking in the country. Key aspects of this plan include prohibiting smoking in public areas such as beaches, parks, and forests, as well as in the vicinity of schools. Additionally, the government aims to phase out single-use disposable vapes.
Another major step is the planned increase in cigarette taxes, which is expected to raise the price of a pack to €13 (about $14) by 2026, up from the current price of €11 ($12). Highlighting the shift in public health policy, the minister emphasized that non-smoking areas will become the standard going forward. These measures are part of France's broader effort to significantly reduce tobacco consumption and its associated health risks.

AWS AI Testing:

Amazon has introduced a new service for companies to better evaluate AI models before deploying them. Announced at the AWS re:Invent conference by AWS Vice President Swami Sivasubramanian, the service, called Model Evaluation on Bedrock, is currently available in preview on the Amazon Bedrock platform.
This service aims to address the challenge developers face in selecting appropriate AI models for their specific needs, like question-and-answer applications or size constraints for a project. Model Evaluation on Bedrock offers two types of evaluations: automated and human. The automated evaluation allows developers to test AI models against various metrics such as accuracy, robustness, and toxicity, using either provided test datasets or their own data.
This evaluation includes models from third parties like Meta, Anthropic, and Stability AI. The human evaluation aspect involves either an AWS human evaluation team or a team chosen by the customer, focusing on additional qualitative aspects like empathy and friendliness, and which automated systems might miss.
Vasi Philomin, AWS Vice President for Generative AI, emphasized the importance of understanding model performance to ensure they meet responsible AI standards and suit the company's specific needs. Currently, AWS is only charging for model inferences used during the evaluation phase of the service. The goal of Bedrock's benchmarking is to help companies measure a model's impact on their specific projects, rather than offering a broad evaluation of the models.

Statistic:

Largest technology companies by market capitalization: