Zephyrnet Logo

One AI image needs as much power as a smartphone charge

Date:

AI in brief Using a text-to-image model to craft an AI-generated image can require almost the same amount of power as that required to charge a smartphone, according to recent research.

In a paper released on arXiv last week, a team of researchers from Hugging Face and Carnegie Mellon University calculated the amount of power AI systems use when asked to perform different tasks.

After asking AIs to perform 1,000 inferences for each task, the researchers found text-based AI tasks are more energy-efficient than jobs involving images.

Text generation consumed 0.042kWh while image generation required 1.35kWh. The boffins assert that charging a smartphone requires 0.012kWh – making image generation a very power-hungry application.

“The least efficient image generation model uses as much energy as 950 smartphone charges (11.49kWh), or nearly one charge per image generation,” the authors wrote, noting the “large variation between image generation models, depending on the size of image that they generate.”

The authors also measured carbon dioxide created by different AI workloads. As depicted in the graphic below, image creation topped that chart

screenshot_graph

Click to enlarge

You can read the full paper here [PDF].

Microsoft invests £2.5B in UK datacenters

Microsoft will spend £2.5 billion to expand its datacenters in the UK and plans to fill them with over 20,000 GPUs by 2026.

The deal is reportedly the single largest investment the tech behemoth has made in Blighty since it launched operations on that side of the pond 40 years ago. Microsoft currently has datacenters in London, Cardiff, and may support other sites in northern England in the future.

Microsoft also pledged to train a million people to work in AI with a multi-million-pound investment. It will provide support for new programs to teach workers skills to help them understand applications and build the technology and plans to establish the first Professional Certificate on Generative AI.

UK prime minister Rishi Sunak announced the investment last week. “Microsoft [is] one of the founding fathers of modern technology and today’s announcement is a turning point for the future of AI infrastructure and development in the UK,” he declared in a statement.

“Microsoft is committed as a company to ensuring that the UK as a country has world-leading AI infrastructure, easy access to the skills people need, and broad protections for safety and security,” added the software giant’s vice chair and president Brad Smith.

Sports Illustrated slammed for slapping AI-generated bylines on AI-generated copy

American sports mag, Sports Illustrated, came under fire last week for publishing AI-generated stories under fake bylines that included AI-generated photos and made-up bios for journalists that don’t exist.

The folks over at science and tech publication, Futurism, thought something looked fishy and traced the image of the proposed journalist writing for Sports Illustrated back to a website selling AI-generated headshots.

When they reached out to the Arena Group, the media biz behind Sports Illustrated, the AI-generated articles and bylines were suddenly taken down. A statement on X released by the sports mag claimed that all the stories were written and edited by humans hired by an external content marketing company AdVon Commerce.

But it decided to spin up fake AI-generated journalism profiles “to protect author privacy.”

Confusingly, however, it also claimed that the content was created and curated using “counter-plagiarism and counter-AI software.” Sports Illustrated announced it has ended its partnership with AdVon and is investigating the issue.

Human writers working at Sports Illustrated were reportedly furious. In a meeting with staff, top executives, including the mag’s editor in chief Steve Cannella, and his boss, the CEO of the Arena Group, Ross Levinsohn, called it “a one-time mistake with an AI company,” according to The Washington Post.

It’s not too surprising that Sports Illustrated turned to spinning out AI-generated content. Another of its titles, Men’s Health, also published a botched article written by a machine that contained factual errors.

Dell to build huge AI compute cluster for unicorn startup

Dell has scored a huge $150 million cloud deal to support Imbue – an AI startup focused on building smart agents that can reason.

The contract was announced last week and shows that it’s not just the biggest cloud providers that can snag prominent AI customers. Anyone with enough GPUs that can be utilized efficiently to support large workloads can do it. In September, Imbue raised $200 million in its Series B round, at a valuation over $1 billion.

The upstart will have access to a high performance computing cluster made up of Dell’s PowerEdge XE9680 servers, which contain two fourth generation Intel Xeon Scalable processors, each containing 56 CPU cores as well as 8 Nvidia H100 and A100 GPUs. The exact number of servers in the cluster was not disclosed.

The upstart’s CTO, Josh Albrecht, revealed that Imbue will use the computers to build new foundational models. “Building a new generation of foundation models requires the very best IT infrastructure, and Dell Technologies has helped us deploy a custom cluster much more quickly than other providers could have,” he declared. The cluster will be managed by a third party – Voltage Park, an AI infrastructure provider.

All cloud providers – especially the big ones like AWS, Google Cloud, or Microsoft – are struggling to meet demands for AI compute. Dell may be able to support Imbue, but that doesn’t necessarily mean it can do the same for other customers. On its earnings call last week, the hardware giant admitted customers ordering its PowerEdge XE9680 servers will have to wait 39 weeks before their gear arrives. ®

spot_img

Latest Intelligence

spot_img