The reason why this newsletter is coming to your inboxes a day late is that we wanted to include the developments from the Samsung Galaxy Unpacked event. Let’s get into it.

First off, it's great that Samsung got straight to the point with AI announcements, in comparison to Apple's WWDC, which took over an hour. Also, within 10 minutes of starting the event, they also addressed the elephant in the room–the Galaxy Ring.

“We’re putting the ability to control your health right into your hands… literally,” said Matthew Higgins, Head of Digital Health Development at Samsung, just as he revealed the ring.

Shaped up to be the next big thing in the health industry, the Galaxy Ring looks promising, offering solid features, though it is somewhat limited compared to the Oura Ring. However, it only works with Samsung devices. As an Apple user, that makes me sad.

The ring uses an AI algorithm to help you understand your sleep patterns and build better habits. It will be available in stores July 24 onward and comes with a $399 price tag.

I couldn't help but feel that Samsung, once a leader in design and innovation, is now closely following popular trends. The company unveiled the Watch Ultra, which resembles the Apple Watch. While it draws inspiration from its competitor, some may find the design less appealing.

Similarly, the Buds3 and Buds3 Pro bear a strong resemblance to AirPods. The Fold6, though new, is quite similar to the Fold5 in terms of display sizes, cameras, and battery, with the main difference being a higher price.

Google announced new features for these devices, including updates to its Gemini AI for Android app, optimized for foldables with overlay and split-screen interfaces. You can access Gemini AI by swiping the screen corner or saying "Hey Google."

The overlay will also allow users to ask questions about YouTube videos. Additionally, Google’s “Circle to Search” feature now supports symbolic math equations, barcode, and QR code scanning and will be available on more devices later this year. 

Samsung aims to boost device sales with these AI features, but Apple’s upcoming AI (which they call Apple Intelligence) could attract Samsung users to iPhones.

What has Google been up to?

Google made headlines when it revealed in its latest environmental report that emissions at the company have skyrocketed nearly 50% over five years due to AI. The annual report shows Google’s progress, or lack thereof, toward its green goals last year. 

The Alphabet Inc. unit revealed it emitted 14.3 million metric tons of carbon dioxide equivalent in 2023—a whopping 48% jump from 2019 and 13% more than 2022. 

Blaming higher energy use at data centers and supply chain emissions, Google admitted that AI might just be its kryptonite in the fight against emissions. Does Google’s goal of carbon-free energy by 2030 seem like a distant dream? You can read more about it here.

Also, Google has updated its political advertisement policy ahead of the US presidential elections. Now, Google will auto-generate disclosures whenever political advertisers mark their election ads as "synthetic or digitally altered."

Previously, Google required political advertisers to add their own clear labels on AI content. The disclosures will show up in feeds, YouTube Shorts, and in-stream ads on mobile, computers, TVs, and the web. For other ad formats, advertisers still need to add their own labels.

OpenAI highlights

Both Apple and Microsoft are dropping their observer seats on OpenAI’s board. Last November, Microsoft took an observer role following Sam Altman's ouster. Similarly, it was reported last week that Apple will also assume an observer position.

Bloomberg reported that the decision comes as antitrust regulators intensify their scrutiny of Big Tech's influence over AI. Microsoft clarified that they are confident in the direction OpenAI is moving towards and are satisfied with their progress, so they believe they no longer need to be a part of the board.

Moving on. I believe that one day, I will write this newsletter without mentioning that someone, somewhere, has sued OpenAI or vice versa, but that day isn’t today.

OpenAI and Microsoft have been sued by the Center for Investigative Reporting (CIR), one of the oldest nonprofit newsrooms in the US, for copyright infringement and depriving the organization of “subscription, licensing, advertising, and affiliate revenue, as well as donations from readers.”

“When they populated their training sets with works of journalism, Defendants had a choice: to respect works of journalism, or not. Defendants chose the latter,” said the complaint uploaded by Mother Jones.

What else made it to the news?

OpenAI’s ChatGPT Mac app storing conversations in plaintext

A software engineer has uncovered a hiccup with OpenAI's ChatGPT app for Mac: It has been apparently saving chats between the user and the bot in plaintext. 

We announced in our newsletter that the ChatGPT Mac app hit the scene a week ago for all users. But then Pedro José Pereira Vieito dropped a bombshell on Threads, revealing that the popular chatbot's desktop app was storing conversations in plain text on your local storage. 

Elon Musk announces Grok 2 in August

Of course, it wasn’t an official announcement. Musk announces everything on his platform X. And Grok 3 will be here by the end of the year, he said in another post.

ElevenLabs voice isolator to take on Adobe?

ElevenLabs, known for voice cloning and text-to-speech models, has launched an AI Voice Isolator. Creators can use it to remove unwanted ambient noise from films, podcasts, or YouTube videos. 

While many providers, including Adobe, offer similar tools, the effectiveness of Voice Isolator remains to be seen.

Note: I tried to use the feature, signed up on the platform and everything, but it wouldn’t even let me click on the ‘Isolate Voice’ button. I followed the instructions as given on their website. I don’t know what I did wrong there.

WhatsApp to reportedly roll out personalized AI avatar generator

WhatsApp is reportedly developing a new generative AI feature allowing users to create personalized avatars for any setting. The feature, spotted in WhatsApp Beta for Android 2.24.14.7 by WABetaInfo, uses user images, text prompts, and Meta’s AI Llama model to generate avatars.

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

• FBI busts Russian AI-driven 'bot farm' spreading fake news in US

  The Justice Department of the US said it had seized two domain names and 968 social media accounts used by Russian actors that spread disinformation in the United States and abroad.

• Moroccan Kenza Layli wins world’s first ‘Miss AI’ beauty pageant
  Kenza Layli, a hijab-wearing AI creator and activist from Morocco, won Miss AI, the world’s first AI beauty pageant.

• OpenAI ban: China gets brand new AI claimed to rival GPT-4 power
  Chinese AI firms, including Baidu and Zhipu AI, seize the opportunity with free tokens and migration services following OpenAI’s China ban.

• Ammo ATM? AI-powered bullet vending machines introduced in US
  The company, according to its website, also says that its vending machines are as easy to use as an ATM.

• Australia launches one of world’s most powerful AI supercomputers
  The system features NVIDIA H100 Tensor Core GPUs with 94GB of high-bandwidth memory, enhancing deep learning, machine learning, and AI applications.

MUST-READ

Mind-reading AI turns thoughts into pictures with unprecedented accuracy

Researchers have achieved a significant breakthrough in the field of neuroscience and artificial intelligence (AI) by successfully reconstructing images from brain activity.

By employing an improved mind-reading AI system, they could produce “closest” reconstructions of images, particularly when using direct recordings of brain activity.

AI PICTURE OF THE WEEK

Researchers at Radboud University in the Netherlands have successfully reconstructed images from brain activity. Below are the original stimuli provided to their subjects, alongside the reconstructed images created using an enhanced AI system.

The researchers achieved these accurate reconstructions of images using direct recordings of brain activity. While it’s not exactly an exact replica of the stimuli image, it comes remarkably close to the real thing. 

You can read more about the mind-reading AI system here.

AI TOOL OF THE WEEK

I found a new playful tool while fooling around on the internet. It’s called ‘Face To All,’ and it simplifies turning your photo into various artistic styles, transforming your face into stylized portraits.

Here’s a tutorial to get you started:

1. Getting to the website

• Visit the website.

• Once you’re on the landing page, you’ll see a button ‘Try Face To All Now’.

• Click on the button.

2. Uploading your picture

• You will see an interface with spaces to upload your picture, pick a style from the gallery, and describe the subject of your photo.

• Once you do all of that, you’ll click on the ‘Run’ button.

3. Generating the image

• The tool will take anywhere from 30 to 70 seconds (or longer) to generate an image for you in the style you chose.

• You also have the option to download the image free of charge.

I personally tried it and found it to be a goofy tool. I gave it Angelina Jolie’s image and chose the ‘ukyio-e’ style, which is a traditional Japanese art style featuring woodblock prints depicting scenes of daily life, nature, and entertainment.

Here are the results:

Pretty neat, eh?