AI is contributing to a rise in energy demand; your child’s next toy could be powered by ChatGPT; AI will be a powerful tool for the future of work; will AI ever outsmart us?

Microsoft doubles down on smaller generative AI; AI's next debate is over whose values it should hold; AI will transform life in emerging economies; a Chinese startup is winning the open source race

Jan 26, 2024

∙ Paid

In the United States, the recent boom in AI adoption, the construction of new factories, and the demand for electric vehicles are causing a significant increase in electricity consumption, which is poised to triple from 2022 levels to as much as 390 terawatt hours by the end of the decade. This surge has caught electric utilities and regulators off guard according to a Bloomberg report, leading to the postponement of plans to close coal-fired power plants in some areas or proposals to turn on diesel-powered generators to provide back up power during power shortages.

Source: Grid Strategies analysis of Federal Energy Regulatory Commission filings

But it’s not all doom and gloom for the environment. A closer look at the energy market shows a promising dynamic at play. Five years ago, for every one US dollar invested in fossil fuels, the same amount was invested in clean energy. Today, the picture looks significantly better: for every dollar spent on generating electricity with coal or methane gas, 1.8 US dollars are spent on investments in renewable energy. And in 2023, solar energy has attracted more capital than global oil production for the first time.

Technology companies have spent the last decade working to ensure their value chain runs on renewable energy, from server to pixel. Most hyperscalers have either achieved net zero emissions in their Scope 1 and 2 emissions or are on a path to achieve net zero by next year, meaning most of the AWS, Azure or Google Cloud infrastructure required by AI workloads is powered by renewable energy. For example, 100% of the emissions required to train Meta’s Llama 2 models were directly offset by the company’s sustainability program, and because Meta openly released these models, the pretraining costs will not need to be incurred by others.

CO2 emissions during pretraining of Meta’s Llama 2 model

Amazon, Google, Meta and Microsoft are also among the largest buyers of renewable energy in the world which in turn incentivizes energy companies to pour even more investments in solar or wind projects.

And the good news doesn’t stop there: The Information reported this week that Microsoft has created a new generative AI team to focus on developing smaller language models that are similar in performance to LLMs like OpenAI’s GPT-4 but use less computing power. French startup Mistral is working on new techniques to get billion parameter-scale models closer in performance to trillion-sized equivalents which will significantly reduce the energy needed to train and run them. And companies such as Qualcomm are already demonstrating LLMs or image generators that can run on a mobile device which will lead to savings in energy usage but also better data privacy protections.

In a recent paper, researchers from the Mohamed bin Zayed University of Artificial Intelligence (MBZUAI) described a new approach that could change the game even further for the energy efficiency of AI models: spiking neural networks (SNNs). The human brain is a marvel of efficiency, capable of complex calculations and learning with minimal energy consumption. But our current crop of large models that power chatbots or generate images and videos require vast amounts of energy to function. One query to ChatGPT uses the equivalent of 3.96 watts, or a third of the battery capacity of an iPhone 13 Pro. Meanwhile, an adult human brain consumes slightly more than a third of a watt of energy in an entire day day, or approximately 8% the energy of one question posed to a large language model.

Inspired by the brain's biological structure, SNNs process information using discrete "spikes" of electrical activity, similar to how neurons communicate. This is a stark contrast to traditional artificial neural networks, which rely on continuous calculations.

Illustration showing the working flow of SNNs

This shift has the potential to dramatically reduce the energy footprint of AI, making it more sustainable and paving the way for exciting new applications including:

Long-range search and rescue: Imagine drones powered by SNNs, able to navigate disaster zones and locate survivors for extended periods without draining their batteries.
Prosthetics: SNNs could lead to more intuitive and natural-feeling prosthetics that better integrate with the human nervous system.
Edge computing: SNNs' low-power requirements make them ideal for running AI tasks on devices with limited resources, like smartphones and wearables.

While SNNs are still in their early stages of development, the potential rewards are vast. By emulating the brain's efficiency, we could unlock a new era of AI that is not only powerful but also sustainable.

And now, here are this week’s news:

❤️Computer loves

Our top news picks for the week - your essential reading from the world of AI

Test Yourself: Which Faces Were Made by AI? [New York Times]
Your Child’s Next Playmate Could Be An AI Toy Powered By ChatGPT [Forbes]
When Might AI Outsmart Us? It Depends Who You Ask [Time]
This Chinese Startup Is Winning the Open Source AI Race [Wired]
Microsoft Forms Team to Make Cheaper Generative AI [The Information]
Why AI needs to learn new languages [The Economist]
AI Should Be a Tool, Not a Curse, for the Future of Work [New York Times]
How AI is reshaping Google’s search page [Fast Company]
AI's next fight is over whose values it should hold [Axios]
There’s an AI ‘brain drain’ in academia [TechCrunch]
Could AI transform life in developing countries? [The Economist]
AI puts every CEO on the hot seat [Axios]
The AI-Fueled Future of Work Needs Humans More Than Ever [Wired]

⚙️Computer does

AI in the wild: how artificial intelligence is used across industry, from the internet, social media, and retail to transportation, healthcare, banking, and more

AI is helping school districts navigate bus driver shortages [Axios]
That Spotify Daylist That Really ‘Gets’ You? It Was Written by A.I. [New York Times]
Travel companies are using AI to better customize trip itineraries [Fortune]
Etsy will give shoppers AI gift suggestions [TechCrunch]
AI to drastically cut time to develop new battery materials, say executives [FT]
AI face-scanning technology to be rolled out at supermarkets to check age of shoppers [The Telegraph]
Nvidia’s RTX GPUs can now upgrade SDR content to HDR using AI [The Verge]
AI is changing the healthcare industry and could transform the way patients get drugs and treatments [Business Insider]
Google’s new Gemini-powered conversational tool helps advertisers quickly build Search campaigns [TechCrunch]
NTT Data to develop AI that can detect driver's cognitive decline [Nikkei]
Humanoid robots will join BMW's production line [Axios]
Ello is using AI to write hundreds of e-books to help kids learn to read [Fast Company]
Sevilla FC Transforms the Player Recruitment Process with the Power of IBM watsonx Generative AI [IBM]
New AI tech could make video-game NPCs a lot more interesting [Fast Company]
Artificial intelligence to analyse rush hour traffic [BBC]
Google Chrome's new AI feature can now organize your browser tabs [Business Insider]
AI’s latest task is helping supermarkets safely unload soon-to-expire food [Bloomberg]
Avatars, robots and AI: Japan turns to innovation to tackle labour crisis [FT]
How $21 billion Balyasny is building an AI equivalent of a senior analyst [Business Insider]
How satellite images and AI could help fight spatial apartheid in South Africa [MIT Technology Review]
How AI Can Find the Perfect Movies, TV Shows and Books for You [WSJ]
Can the power of artificial intelligence be harnessed to help predict Australia’s weather? [The Guardian]
Want Your Packages Delivered Faster and Cheaper? AI Is on the Case [WSJ]
You can use AI in Instagram to generate fun backgrounds for your stories. Here's how [ZDNet]

🧑‍🎓Computer learns

Interesting trends and developments from various AI fields, companies and people

Keep reading with a 7-day free trial

Subscribe to Computerspeak by Alexandru Voica to keep reading this post and get 7 days of free access to the full post archives.