📷 World’s first ‘Poetry Camera’ uses AI-power to turn photos into poems, Mark Zuckerberg on energy constraints limiting AI growth

Over the last week, AI's energy use and the restrictions it poses have come under the spotlight. In a podcast with Dwarkesh Patel, Mark Zuckerberg confirmed that energy constraints had become the largest bottleneck to building AI.

Zuckerberg stated that energy constraints will soon bind AI's growth, not chips. 

"Over the last few years, there's been this issue of GPU production," he said. "Even companies that had the money to pay for the GPUs couldn't necessarily get as many as they wanted because of all these supply constraints.”

"Now I think that's getting less, and you're seeing a bunch of companies thinking 'wow we should really just invest a lot of money in building out these things.' I think that will go [on] for some period of time."

With AI models requiring ever greater investment to improve "there is a capital question of at what point it stops being worth it to put the capital in," Zuckerberg said. "But I actually think that, before we run into that, we're going to run into energy constraints."

He said that while software is only "somewhat regulated" the energy industry is a heavily regulated sector. "When you're talking about building large new power plants or large build-outs and building transmission lines that cross public or private land... you're talking about many years of lead time."

What would Meta do with 10x more money? More energy, which would allow bigger clusters, but the true bottleneck is time. Energy gets permitted at the speed of government and then has to be physically built.

"If we wanted to stand up some massive facility, to power that is a very long-term project. I think [some people will] do it, but I don't think this is something that can be quite as magical as 'you get a level of AI, get a bunch of capital, and put it in [a big data center]."

"I think we would probably build out bigger clusters than we currently can if we could get the energy to do it," Zuckerberg said.

"No one has built a 1GW data center yet. I think it will happen. This is only a matter of time but it's not going to be next year."

Did a friend forward this e-mail to you?

IE₊ SUPPORT INTERESTING ENGINEERING
_{Invest In Science And Engineering}

Enjoy exclusive access to the forefront of AI content, highlighting trends and news that shape the future. Join a community passionate about AI, delve into the latest AI breakthroughs, and be informed with our AI-focused weekly premium newsletters. With IE+, AI reporting goes beyond the ordinary - and it is Ad-Free.

NEWS

• 🩸 AI-powered tool detects cancer in minutes with one drop of blood

  Scientists in China have developed a test that detects three types of cancers with just one spot of dried blood.
  

• 📹 Meta’s Llama 3-powered AI assistant rolls out across FB, Insta, and WA

  Built on Meta’s latest large language model called Llama 3, the AI assistant is Meta’s rival to ChatGPT and Google’s Gemini.
  

• 🧠 World’s largest neuromorphic computer by Intel works like human brain

  Powering Hala Point is Intel’s Loihi 2 neuromorphic processor that represents a significant upgrade over its predecessor, Pohoiki Springs.

MUST READ

World’s first ‘Poetry Camera’ uses AI-power to turn photos into poems

The latest in the never-ending series of artificial intelligence (AI) innovations is one for the artists, those who love the written word. Dubbed “Poetry Camera”, the AI-enabled device stays true to its name—it turns pictures into poems!

Kelin Carolyn Zhang and Ryan Mather collaborated on the Poetry Camera. It is an open-source project that merges technology with artistic flair, yielding a creation that transcends boundaries in both domains.

The Poetry Camera outwardly resembles a typical Polaroid camera. However, rather than capturing photographs, it transforms the visual input it receives into evocative poems that vividly describe the shot that was captured.

AI PICTURE OF THE WEEK

TUTORIAL

Give AI copy some style

AI copywriting is invariably generic and easily identifiable, but did you know you can prompt AI to write according to editorial style guides? Chat GPT understands style guides such as the AP and Economist style guides. Moreover, it knows the rules of the copy editor's favorite, Strunk & White's 'Elements of Style'.

Next time you ask AI to write some copy, append your prompt with 'write to the Economist style guide and apply Strunk & White's 'Elements of Style' and see the difference.

TOOLS OF THE WEEK

🗣️ Adobe Photoshop: Adobe is adding new generative AI tools to Photoshop, powered by Adobe’s new Firefly Image 3 foundation model. According to Adobe, it will give users more control and personalization over the styles of images they generate, enable the use of reference images and deliver new levels of photographic quality with better lighting, positioning and more variety, alongside a better understanding of text prompts and scenes. These new tools are available today via the Photoshop beta desktop app, and will be generally available “later this year”.

🎵 Suno: Suno have released their V3 music generator that they claim is capable of producing radio-quality music. V3 enables you to make full, two-minute songs in seconds and is now available to all users.

🧐 Semantic Scholar: This is a free, AI-powered research tool for scientific literature. It indexes more than 2 million academic research papers and can extract key findings from them in seconds. It can also help you cite and find influential literature, and will notify you of new papers or citations of your interest.