🤹🏻‍♂️Meta's Latest Innovation

AI's energy appetite pushes Google's emissions up 50% in five-year span

Key Points 

• Tech giants' greenhouse gas emissions have surged: Google's by 50% in 5 years and Microsoft's by 30% since 2020, largely from expanding data centers.

• AI's rapid growth challenges corporate climate goals and threatens to worsen environmental impacts.

😷News - Google's greenhouse gas emissions have increased by almost 50% in the last five years, largely due to the energy-intensive requirements of data centers supporting artificial intelligence. The company's annual environmental report for 2023 also reveals a 13% increase in emissions compared to the previous year, totaling 14.3 million metric tons.

The tech giant admitted that achieving its "extremely ambitious" goal of reaching net zero emissions by 2030 will be challenging. It cited significant uncertainties, particularly concerning the unpredictable environmental impact of AI.

Furthermore, the company acknowledged that as it integrates AI more extensively into its products, reducing emissions may prove difficult due to the increased energy demands of AI computing and the emissions linked with expanding its technical infrastructure investments.

🤕What's more? In parallel, Microsoft, which has committed to becoming carbon-negative by the end of this decade, reported a nearly 30% increase in emissions since 2020, also largely due to the construction of data centers.

🫠Why this is problematic - The International Energy Agency predicts that by 2026, data centers could double their electricity consumption to 1,000TWh, equivalent to Japan's current electricity demand. Research from SemiAnalysis suggests AI could drive data centers to consume 4.5% of global energy by 2030.

Additionally, AI's rapid advancement raises significant concerns about water usage. One study estimates it could consume up to 6.6 billion cubic meters by 2027 — nearly two-thirds of England's annual consumption. Thus, AI is not only pushing corporate climate goals further out of reach but also appears poised to exacerbate the situation.

Meta's latest AI model converts text to 3D images in under 60 seconds

Key Points 

• The 3D Gen model, under development, supports high-resolution textures, material maps, PBR, and generative re-texturing.

• Initial 3D model creation with Meta's 3D AssetGen takes about 30 seconds, with texture refinement via text prompts adding another 20 seconds.

🧙🏻‍♂️News - Meta recently unveiled a new AI image generation tool called "3D Gen,"  that can convert text into high-quality 3D images in less than a minute. This system can also enhance both AI-generated and artist-created images by applying new textures and skins based on text prompts.

A study by Meta's Gen AI research team revealed that the 3D Gen model will offer high-resolution textures, material maps, and support for physically-based rendering (PBR) and generative re-texturing. 

The team estimates that creating an initial 3D model using Meta’s 3D AssetGen model takes about 30 seconds. Users can then refine or replace the model's texture via text prompts using Meta 3D TextureGen, which should take an additional 20 seconds.

👾Significance & impact - Although the system is still under development and not yet available for public use, the technical advancements could revolutionize various creative fields, including game and film effects and VR applications. By enabling users to quickly and intuitively create and edit 3D-generated content, it could significantly reduce the barrier to entry for these activities, potentially having a major impact on game development.

Apple secures observer role on OpenAI board as part of AI pact

Key Points 

• Phil Schiller, Apple's App Store chief and former marketing head, has been selected for the role. 

• Apple's observer position would match that of Microsoft. This may complicate discussions on AI initiatives involving both Microsoft and OpenAI, given the historically complex relations between Apple and Microsoft.

📋Context of the news - At WWDC 2024, Apple announced a partnership with OpenAI. It was later revealed that although the collaboration will bring OpenAI's technology to millions of iPhone users, Apple won't be cutting a check to OpenAI. 

🍎News - Now, according to a new update reported by Bloomberg, Apple will also assume an observer role within OpenAI as part of their agreement.

The report notes that Phil Schiller, Apple's App Store chief and former marketing head, has been selected to take on the role of board observer. In this capacity, Schiller will be able to attend board meetings but will not have voting privileges. This arrangement is set to officially begin later this year, with Schiller yet to participate in any meetings.

🪑Potential implications - Apple joining the OpenAI board could potentially complicate discussions on AI initiatives involving Microsoft and OpenAI at board meetings, especially since this role will place Apple in a comparable position to Microsoft, a company with which it has had complex relations in the past.

That being said, it is customary for board members to recuse themselves from sensitive discussions as needed. Nevertheless, it will be interesting to see how this plays out.