Starting with the latest in AI this week…

Figure launches 'most advanced' humanoid robot

The OpenAI-backed company that wants to build humanoid robots for commercial purposes has unveiled its new AI-driven robot, Figure 02. 

Founded by entrepreneur Adcock in 2022, the company introduced its first robot, Figure 01, last year and secured funding from major investors like Nvidia, Microsoft, Intel Capital, and Jeff Bezos's private fund. 

As competition in humanoid robotics intensifies, Figure plans to incorporate advanced AI models, including OpenAI's conversational AI, into its robots for enhanced communication and interaction capabilities. Will we soon have a walking-talking robot?

The 2-minute long video shows Figure 02 at work. The 5’6” tall robot, with a run time of only five hours, can be seen at work–assembling and placing car parts at the BMW factory. But the robot’s gait has caught people’s attention the most, it seems.

The robot can be seen walking around taking teeny tiny steps. One YouTube user's comment read: “Getting them to walk like they've soiled their pants is genius - makes them less threatening.”

Speaking of what caught people’s attention last week…

…A wearable AI necklace that promises constant connection 

With a sleek, rounded design and a glowing central element, Friend is an AI activated device that listens to your conversations, even if you’re not directly talking to it 👀 and responds via text without being prompted. Want a direct chat? Just press the center of the device and start talking.

Check it out:

The device, available only in white, features a Bluetooth microphone and sends data to the cloud for processing. The company assures users of end-to-end encryption and promises not to store any audio or transcripts long-term. It’s priced cheap at $99. 

Some may find the idea of continuous audio recording off-putting, but Avi Schiffmann, its creator, believes it's crucial for achieving a natural, engaging experience. 

But there’s a bigger storm brewing for the creators of Friend with another person staking claim as the original creator of the device. More about that in our AI Picture Of The Week section.

Stocks crash for AI-heavy investors: Nvidia, Apple, Microsoft, Tesla hit hard

Monday was not a great day for AI, as the global stock market crash resulted in significant losses for major US technology companies, which had been riding on the AI boom. The waves are breaking…

The "magnificent seven" tech giants—Alphabet, Amazon, Meta Platforms, Microsoft, Tesla, Apple, and Nvidia—suffered their sharpest decline since the start of the COVID-19 pandemic, wiping out nearly $900 billion in market value. 

Apple and Nvidia were hit hardest, with losses of approximately $270 billion and $380 billion, respectively. The downturn followed a weak jobs report, fueling fears of a US recession. 

Additionally, Warren Buffett's Berkshire Hathaway announced it had halved its stake in Apple, further shaking investor confidence. The market was also impacted by delays in Nvidia’s highly anticipated Blackwell chips due to design flaws, exacerbating concerns about the tech sector. 

Chip stocks tumbled, with the Philadelphia Semiconductor Index falling nearly 3%, and shares of companies like Advanced Micro Devices, Intel, Super Micro Computer, and Broadcom dropping by up to 7.8%.

Meta is offering celebrities millions to clone their voice for AI projects

After launching AI Studio last week, Meta is in the process of negotiating with Hollywood celebrities like Judi Dench, Awkwafina, and Keegan-Michael Key to use their voices for AI projects, potentially for a chatbot or digital assistant, reported Bloomberg. 

The deals, aimed at Meta’s upcoming Connect 2024 event in September, involve significant payments but face challenges over usage terms. Meta wants broad rights for the voices, while talent representatives seek more restrictive terms. 

Despite concerns from the creative community about AI’s impact on jobs, SAG-AFTRA has reached an agreement with Meta. Previously, Meta had similar celebrity voice projects but discontinued them.

Video game performers strike over AI concerns

We haven’t heard the last of creators’ protests against AI. Video game actors, represented by SAG-AFTRA, have gone on strike alongside their Hollywood counterparts, protesting the use of AI by game producers. 

The strike, which began on July 26, 2024, has halted work at major companies like EA, Activision, and Epic Games. The actors are demanding protections against the use of AI to replicate their likenesses and performances without consent or fair compensation.

Character.AI co-founders return to Google

In a surprising turn of events, Character.AI co-founder and CEO Noam Shazeer is heading back to Google after leaving in October 2021 to start the a16z-backed chatbot startup. 

Shazeer, who previously led the team that developed Google's LaMDA (Language Model for Dialogue Applications), will bring his expertise back to the tech giant. He will be joined by Character.AI co-founder Daniel De Freitas and several other startup employees.

AI Music generators’ bold respond to Sony, UMG, and Warner

AI music generators Suno and Udio have responded to copyright lawsuits from major recording labels, arguing that their use of copyrighted works is legal and falls under ‘fair use.’ 

The lawsuits, filed by labels like UMG, Sony, and Warner, allege that Suno and Udio copied and used decades of popular recordings to train their AI models, potentially displacing human artistry. 

Shocker: Microsoft lists OpenAI as a competitor

Whoa, whoa, whoa… what now? Valley friends turn frenemies? Microsoft has labeled OpenAI, a company it heavily invested in (all $13 billion of it) as a competitor in its latest annual report. 

This highlights a complex relationship, as both companies offer competing AI products despite their partnership. Microsoft integrates OpenAI’s AI models into its products, and competes with the ChatGPT maker in areas like AI chatbots and cloud services. 

Despite this, both companies acknowledge competition as part of their partnership. Recent events, including the temporary ousting of OpenAI's CEO and Microsoft's appointment of a new AI unit leader, indicate potential internal competition and strategic shifts.

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

• World’s 1st AI-powered hearing aids boost speech understanding by 53 times
  Sonova, a global leader in hearing aid technology, announces the release of Phonak Audéo Sphere™ — the first platform that integrates AI and dual-chip technology.

• In a first, gen AI LLM deployed aboard International Space Station's lab
  Generative AI in space is truly the new frontier and this capability unlocks the potential for on-orbit generative AI to integrate and develop mission-critical solutions.

• AI Alan Turing: World-1st interactive hologram of WWII British codebreaker
  The AI character, a world first, will tailor its responses to different audiences, including individuals, groups, and children.

• Caltech’s damaged robot adapts to swim like injured fish with AI power
  Bioinspired adaptation techniques allow flapping robots to remain functional and efficient despite significant damage.

MUST-READ

No semen needed: AI detects male infertility from blood with 100% accuracy 

The infertility risk of a man can now be predicted through a basic blood test, and the credit goes to a new tool developed by researchers in Japan.

This advanced method leverages artificial intelligence (AI) to simplify the process for men to determine whether they may face trouble having children.

The AI system analyzes hormone levels from the blood test and can predict male infertility with about 74% accuracy. Importantly, it has demonstrated 100% accuracy in identifying non-obstructive azoospermia, a condition where no sperm is found in the semen.

AI PICTURE OF THE WEEK

Launched on the International Day of Friendship, the device Friend has found itself in an unfriendly battle of who-launched-it-first. On the day Avi Schiffmann launched his device, another founder named Nik Shevchenko released a diss track accusing Schiffmann of stealing his idea. 

In what can be best described as a rap, Shevchenko says he is developing an open-source pendant that looks similar to Friend and is also named…wait for it…drumroll, please… Friend. 

In the rap, Shevchenko claims, "For those who don't know / I built Friend / renamed your Tab/ jacked my style."

But allegedly, it isn't the first time Shevchenko has created an open-source device similar to another AI hardware product. In March, the company Humane reportedly sent him a cease-and-desist order for launching a device called Whomane similar to the Humane AI Pin.

But, this device can yap! What confounded me was one bit from the advertisement in which a woman dropped mayonnaise on the device, and the device responded back with ‘yum.’ 

It got me wondering: can the device ‘see’ as well?

While many appreciate the aesthetics of the trailer (and the actors in it) showcasing how it can be used in day-to-day activities, some reports have called Friend ‘creepy’ or compared its trailer to a Black Mirror episode. And then, of course, there is the meme-fest. One of our favorite memes on the internet was this:

Will it meet the same fate as that of the Humane AI Pin and Rabbit r1? Time will tell.

Also, you guys did well in last week’s AI picture challenge. Around 60% of you correctly identified the real image (which is Image 1), so give yourselves a pat on the back 👏

One subscriber, who went for Image 2, commented that they went for the fake-looking one because AI has gotten really good at creating realistic images, which it sure has. Fully agree with that.

AI TOOL OF THE WEEK

This week, we got to try our hands on an AI tool that has been around for some time but hasn’t been spoken about much. Anyway, it’s a useful tool for podcasters who inadvertently swear in the midst of a podcast or people generating content that they want to keep PG-13 friendly. 

This tool, SwearAway, is pretty self-explanatory from its title. It removes cuss words and swears from an audio file. “SwearAway is here to turn your audio nightmares into a f***ing dream come true,” says the company demo.

I won’t be writing the whole tutorial for it, because the demo video does a pretty good job of explaining how the tool works:

Of course, I had to try it. Since I am a millennial and 50 Cent’s ‘Candy Shop’ lyrics sent shockwaves through my 15-year-old body back then, I decided to test it out. Yes, I am ancient, and yes, there are more explicit songs I could have chosen, but here we are.

I uploaded the explicit version (there is a cleaner version as well) of ‘Candy Shop’ and SwearAway took about 30 seconds to upload, create a transcript and produce a clean version of the song.

Results? Words like “nympho”, “s**t”, “erotic”, and “a**” were highlighted in the transcript, and the tool gave me the option to download the cleaned audio. You can also blank out any additional words that you’d want removed and download the clip. 

The app gives you three options to remove the cuss words: silence them, a 1000Hz Bleep or sidechain. I went with the classic 1000Hz Bleep and the app delivered. Try it out for yourself and let me know if you thought this was useful. See you next week. Cheers!

You can write to me at [email protected] to give any feedback.