The High Energy Consumption of AI Systems
AI and Data Tech Trends - Issue 38

The High Energy Consumption of AI Systems

Thank you for reading my latest article The High Energy Consumption of AI Systems. Here at LinkedIn, Forbes, The AI Journal and Brainz Magazine I regularly write about data and technology trends.??

To read my future articles simply join my network here or click 'Follow'.???

---------------------------------------------------------------------------------------------------------------

AI energy demands could soon match the entire electricity consumption of?Ireland

According to a study by researcher Alex de Vries from Vrije Universiteit Amsterdam, the energy consumption of AI systems could soon rival that of an entire country. The training phase of large language models is already known to be energy-intensive, but the inference phase, where models generate live responses to user queries, may contribute significantly to their overall energy demands. For example, OpenAI's ChatGPT required 3,617 Nvidia HGX A100 servers, consuming an estimated 564 MWh per day. Google also reported that 60% of AI-related energy consumption from 2019 to 2021 came from inference. Although the hypothetical scenario of AI energy demands matching an entire country's consumption is not imminent, it highlights the need for optimizing AI systems and critically assessing their necessity. To read more about the environmental impact of AI systems, click the link below.

Read more: AI Energy Demands Could Soon Match Entire Electricity Consumption

Utah Files Lawsuit Against TikTok, Citing Child Safety Concerns and Ties to China!

Utah has taken legal action against TikTok, alleging child safety issues and raising concerns about the app's ownership by China-based company ByteDance. Attorney General Sean Reyes has described TikTok as an "addictive product" and accused it of misleading users about its connections to China. The lawsuit compares the app to a slot machine, claiming it manipulates users through dopamine triggers. Utah's strict social media laws require parental permission for teenage social media usage. To delve deeper into the details of the lawsuit and the allegations, click the link below.

Read more: Utah Sues TikTok Over Child Safety Issues and Its Links to China

Firefox to Integrate Fakespot's 'Fake Reviews Detector' to Combat Misleading Amazon Reviews!

Mozilla, the company behind Firefox, has acquired Fakespot, a startup known for helping users identify fake news and reviews through a website and a browser extension. Now, Fakespot's Review Checker is set to be integrated into Firefox version 120 for desktop and Android starting from November 21, 2023. This move aims to address the issue of deceptive product reviews on platforms like Amazon, Yelp, TripAdvisor, Walmart, and eBay. By leveraging Fakespot's sophisticated artificial intelligence (AI) and machine learning (ML) system, Firefox aims to empower users to identify and avoid fake reviews that artificially inflate product rankings. To get more details on this development, click the link below.

Read more: Firefox will have a built-in ‘fake reviews detector’ — Amazon is in trouble

Top 3 AI Publications Of The Week

Large Language Models as Analogical Reasoners

The research introduces "Analogical Prompting" for large language models (LLMs). It's a method to help LLMs reason by making them self-create relevant examples or knowledge before trying to solve a task, mimicking how humans use past experiences to address new problems.

Read the paper

Representation Engineering: A Top-Down Approach to AI Transparency

This research introduces Representation Engineering (RepE), a technical approach to enhancing the transparency of AI systems. RepE focuses on understanding and controlling high-level representations within AI models, departing from traditional bottom-up approaches that examine individual neurons and connections.

Read the paper

Efficient Streaming Language Models with Attention Sinks

This research introduces StreamingLLM, a new framework that lets Large Language Models (LLMs) handle extremely long texts without a performance drop. StreamingLLM addresses the challenges LLMs face when applied to streaming applications where lengthy interactions are expected.

Try StreamingLLM

Data Analytics Training

Join Our Training Programs

Our training programs are meticulously crafted to cater to individuals and businesses at every skill level – from beginners to seasoned professionals. Whether you're looking to master data analysis, dive into machine learning, or understand the intricacies of AI implementation, we have a program that suits your aspirations.

Onyx Data Training
Onyx Data Training

---------------------------------------------------------------------------------------------------------------

About Leon Gordon?

Leon Gordon, is a leader in data analytics. A current Microsoft Data Platform MVP based in the UK and founder of Data Analytics and AI consultancy Onyx Data. During the last decade, he has helped organisations improve their business performance, use data more intelligently, and understand the implications of new technologies such as artificial intelligence and big data.?

Leon is a Thought Leader at the Forbes Tech Council and also an Executive Contributor to Brainz Magazine, a Thought Leader in Data Science for the Global AI Hub and The AI Journal, chair for the Microsoft Fabric – UK community group and the DataDNA data visualisation community as well as an international speaker and advisor.

要查看或添加评论,请登录

Leon Gordon的更多文章

  • Power BI Certification

    Power BI Certification

    Hello everyone! Today’s newsletter is about one of my favourite things: Power BI. As both a Microsoft MVP and owner of…

    8 条评论
  • All About Data Algorithms + How to work with them in Power BI!

    All About Data Algorithms + How to work with them in Power BI!

    Hello everyone, Happy Wednesday! Today’s newsletter is all about data algorithms. One of the most important parts of…

    6 条评论
  • 100th Edition of AI and Data Tech Trends

    100th Edition of AI and Data Tech Trends

    Hello everyone! Today we have a very special 100th edition of AI and Data Tech Trends! I wanted to start by saying a…

    4 条评论
  • Big Data Analysis in Power BI

    Big Data Analysis in Power BI

    Hello everyone, Last week’s newsletter was all about data centres and this week we’re talking about what they handle…

    5 条评论
  • How Have Data Centres Become So Valuable and In Demand?

    How Have Data Centres Become So Valuable and In Demand?

    Hello everyone, In today’s newsletter, I’m going to talk about data centres. What are they? Where are they? And - more…

    4 条评论
  • My Favourite Power BI Tips and Tricks

    My Favourite Power BI Tips and Tricks

    Hello everyone! Happy Wednesday. Today I will be sharing some tips and tricks for Power BI.

    2 条评论
  • How to Build a Data Strategy

    How to Build a Data Strategy

    Happy Wednesday! In this week’s newsletter, I will be talking about how to build and execute a successful data strategy…

    1 条评论
  • Joining the Data Industry in 2025

    Joining the Data Industry in 2025

    When I joined the data industry, there was a lot I got wrong. And a lot more I’ve learned along the way since! Here’s…

    3 条评论
  • Exciting AI News & Applications

    Exciting AI News & Applications

    We’ve had a busy start to 2025 with a lot of exciting plans for AI to kick off the year! From potential applications on…

    1 条评论
  • 2025 AI & Technology Predictions

    2025 AI & Technology Predictions

    Happy New Year! What better way to kick off 2025 than with a new edition of AI and Data Tech Trends! This week we will…

    2 条评论

社区洞察

其他会员也浏览了