Daily AI Newsletter: Adobe introduce Firefly 3

In today’s email:

💧 Thanks To Ai, China’s Data Centers Will Drink More Water Than All Of South Korea By 2030
📚 AI for Data Journalism: demonstrating what we can do with this stuff right now
🤔 How many AI models is too many?
🧰 9 new AI-powered tools and resources. Make sure to check the online version for the full list of tools.

Adobe’s new Firefly model makes it easier to use Photoshop’s AI tools

Adobe is integrating new generative AI tools into its Photoshop beta desktop app, enhancing user experience with powerful features like Reference Image, Generate Background, and Enhance Detail. These tools, powered by Adobe's latest Firefly Image 3 model, are designed to streamline creative workflows by using reference images to generate designs that match in style and color, creating new backgrounds for product photography, and sharpening image details for improved clarity.

The introduction of Firefly Image 3 also marks a significant improvement in image generation, providing photorealistic quality and more accurate interpretation of complex text prompts. Adobe is emphasizing the safety and commercial usability of these tools, noting that images used for reference will not train the Firefly model and users must have the rights to any images they use.

Additionally, Adobe is enhancing Photoshop with non-AI tools such as an Adjustment Brush for non-destructive color changes and Adjustment Presets for quick application of filters. A revamped Font Browser improves access to Adobe's extensive font library without leaving the app, demonstrating Adobe's commitment to both cutting-edge AI and traditional image editing tools to cater to a wide range of creative needs.

Turn your ideas into websites effortlessly with AI

Mixo is the fastest and easiest way to launch a product, service, or idea online. Create beautiful websites in seconds. No design or coding skills are required.

Try for free ➜

Meta’s AI Needs to Speak With You

Meta is heavily investing in integrating its AI chatbots across its main platforms, including Facebook, Instagram, WhatsApp, and Messenger. Mark Zuckerberg has updated Meta's AI, promoting its capabilities and stating his ambition towards achieving Artificial General Intelligence (AGI). The chatbots are designed to engage users by responding to comments, participating in group messages, and offering suggestions within the social feeds, thereby keeping users within Meta’s apps instead of navigating elsewhere for information or interactions.

The practical application of Meta’s AI is most evident in its search features across the company's platforms. Users can now conduct searches directly through Meta's chatbots, obtaining real-time information without leaving the app. This setup aims to streamline user interactions by reducing the need to switch between apps or use external search engines, presenting a future where the web is less a navigable space and more a data resource for AI to summarize and reference.

Despite its innovative facade, Meta’s AI has received mixed reactions, illustrating both technological prowess and the challenges of implementing such advanced AI in everyday contexts. For example, when searching for “New York” on Instagram, the results were unexpectedly mundane, highlighting the potential gap between AI capabilities and practical user needs. This kind of deployment suggests Meta’s focus on enhancing user engagement within its platforms, at significant financial and computational costs.

Thanks To Ai, China’s Data Centers Will Drink More Water Than All Of South Korea By 2030

China's AI expansion is leading to a significant increase in the water consumption of its data centers. According to a report by China Water Risk, by 2030, these facilities could use about 792 billion gallons of water, an amount equivalent to the residential water needs of the entire population of South Korea. This surge in usage is due to the extensive water cooling systems required to manage the heat produced by training and maintaining AI models.

The rise in AI technology usage has a global impact, significantly increasing water consumption in data centers worldwide. In the U.S., Microsoft reportedly used 185,000 gallons of water in the training of GPT-3 alone, while Google's total water usage in 2022 reached 5.6 billion gallons. The energy-intensive nature of AI, relying on specialized chips, starkly contrasts with more traditional online search methods in terms of water usage.

Concerns about the sustainability of this trend are growing, with experts like Arm Holdings Plc CEO Rene Haas warning that the energy consumption of data centers could surpass that of India by 2030. There is a pressing need for innovations in energy efficiency, particularly in the development of more efficient AI chips, to mitigate the environmental impact of the expanding digital infrastructure.

Other stuff

AI for Data Journalism: demonstrating what we can do with this stuff right now
Color Naming: Human vs. GPT-4
Netflix Accused of Using AI Photos in True Crime Documentary
Linus Torvalds on Security, AI, Open Source and Trust
How United Airlines uses AI to make flying the friendly skies a bit easier
How many AI models is too many?
The Pentagon is spending billions on Big Tech and Silicon Valley startups as it goes all-in on AI