🍎 Is Apple’s AI branding on point or on the nose?

Welcome to this edition of Loop!

To kick off your week, we’ve rounded-up the most important technology and AI updates that you should know about.

‏‏‎ ‎ HIGHLIGHTS ‏‏‎ ‎

• How Generative AI is being used to train robots

• EU’s huge tariffs on Chinese EVs

• Why Albedo could disrupt the satellite industry

• … and much more

Let's jump in!

1. EU will introduce tariffs of 38% on Chinese EVs

In a move that has been anticipated for months, the European Union announced that it will impose tariffs of up to 38% on electric cars manufactured in China.

The EU has completed an investigation into Chinese electric cars, which found that their entire supply chain was benefiting from government subsidies.

This allowed Chinese automakers to significantly reduce their production costs and gave them an unfair competitive advantage over their European counterparts.

The decision could also slow Europe’s transition away from fossil fuels, as electric vehicles will become more expensive.

German manufacturers - such as BMW, Mercedes-Benz, and Volkswagen - are worried about the decision. They all sell a lot of vehicles in the Chinese market and have significant production and R&D operations in the country.

Many in the industry have warned that this move will lead to higher prices and could spark a trade war with China.

2. Tesla shareholders approve Elon Musk’s $56 billion compensation package 

Musk’s huge compensation package has been given the go-ahead by Tesla’s biggest shareholders.

The deal was heavily reliant on Tesla hitting their financial targets. If this didn’t happen, Musk wouldn’t be paid at all. Although, the sheer size of the CEO’s compensation package isn’t something that we have seen before.

While Tesla’s shareholders have now approved the package, the final ruling still rests with a Delaware judge who previously rescinded it in January.

Tesla and Musk have faced multiple lawsuits this week, including claims of insider trading and diverting resources to Musk's competing AI company, xAI.

3. Mistral AI raises $640 million

The Paris-based AI startup is now valued at $6 billion and follows the deal they struck with Microsoft a few months ago.

The company was founded by former researchers from Meta and Google DeepMind, with the aim of creating AI models that can rival GPT-4, Claude 3 and Llama 3.

They’ve received a lot of attention in the last 12 months, largely due to the fact that they have released open-source models that are very capable.

Mistral have made great progress during that time, but they now face the challenge of attracting corporate customers if they want to increase revenue.

By partnering with Microsoft, Mistral hopes that they can get a better insight into what big companies want to see and then plug those gaps in the market.

4. BeReal has been acquired for $535 million

Voodoo, which is a French mobile apps and games publisher, has acquired the social media platform BeReal for €500 million (equivalent to $535 million).

The app took a different spin on social media, as it encouraged users to be more authentic. Users were prompted to post an unedited image within a two-minute window, which would change over time.

BeReal gained a lot of popularity during the pandemic and already has over 40 million active users. But it has been struggling to grow its user base over the last year and find a clear funding model.

The deal works well for Voodoo, as BeReal is one of the most popular social networks in their 3 core markets - the US, Japan, and France.

It’s hoped that Voodoo’s expertise with other platforms can boost BeReal’s user base and start to generate revenue streams for the app.

5. Generative AI is bringing robots a step closer to “general purpose” tasks

Researchers at MIT have introduced a new method called Policy Composition (PoCo). This is a new way to train robots on how to complete general tasks, through the use of diffusion models.

The researchers trained separate models, each on a specific data source - such as simulation data, robot data, or human videos. These models were then combined later on.

Separating the models is important as it allows us to leverage the strengths of each data source, without having to worry about combining the data beforehand - which can be quite challenging.

What’s interesting is that the robots saw performance increase by 20%. Ultimately, the aim here is to allow robots to swap between different tools and be more flexible in their work.

To date, most of the industry has focused on creating robots that specialise on only one task - think of the robots we often see in factories. They don’t have the flexibility to adapt to changes, which means it can be costly to update them.

A lot of money has been invested in general robotics startups, like 1X and Agility, who are hoping to change that.

Apple brings focus to the AI industry

The company unveiled Apple Intelligence, which aims to provide more personalised experiences for users. But the key thing to stress is that your data will not be used to train new AI models.

Overall, it was quite refreshing to see the range of AI features that were announced - especially since every product in the last 12 months has looked like a chatbot.

Siri was given a significant upgrade and can now integrate with ChatGPT when it doesn’t know how to answer a question.

Students can write mathematical equations in the Notes app and Apple’s AI will write the answer for them. If students use it correctly, it could really help them understand how to solve different equations and learn faster.

Soon you’ll be able to use Apple’s AI-powered writing tools to rewrite, proof-read, and summarise notes in any app. There are also new features that let you create and edit photos, without having to use expensive software.

Quite frankly, there’s too many announcements to simply list here and I really encourage you to check them out online.

Although, one thing that has proved a bit divisive is the name “Apple Intelligence”. Is it a genius way to “own” and rebrand AI, or a bit on the nose?

Personally, I don’t mind what it’s called. The average user will not be saying “Apple Intelligence” 50 times a day and Apple probably doesn’t want them to.

By using this name, they’re clearly communicating to investors that the company is building AI tools - which Apple has come under fire for in the last year.

It also allows Apple to avoid the toxic acronym “AI”, which is often misused by the media and can worry the general public. Essentially, it’s the best of both worlds.

Apple’s done a great job in integrating the technology into their devices. I’d love to see more of them being developed for the Vision Pro, as there’s huge opportunities there - but that might be something for next year.

This humanoid robot can drive cars (well, sort of)

Researchers at the University of Tokyo have developed a humanoid robot called Musashi, which was trained to drive an electric car on a test track.

The robot was equipped with cameras for eyes and mechanical hands and feet, which is pretty strange to see.

Musashi was able to navigate around a corner and obey the traffic lights, but it did so at a very slow pace. This was largely down to an abundance of safety and technical limitations with the robot.

The robot was able to use the accelerator in other tests, but struggled to keep a consistent speed when the road had an incline. Of course, the technology is in the very early stages but it’s an interesting project.

A lot of self-driving companies have focused on building sensors and cameras into the car, with nothing actually in the driver’s seat.

That approach has proved somewhat successful for driving around a city, but hasn’t been perfect as they often drive into wet concrete or block roads - even after a decade of research.

The researchers have suggested that robots could be used as taxi drivers in Tokyo, although I’m not so sure.

While there’s a long way to go with this research, and I’m not sure why it’s really needed, it’s still fascinating to see what we can currently do - even if it’s very, very slow.

🕵️ Former NSA head joins OpenAI’s board

🎙️ Spotify is testing generative AI for voiceover ads

🔒 Hackers stole a “significant volume of data” from Snowflake customers

💸 Amazon says it'll spend $230 million on generative AI startups

🏥 London hospitals cancel 800 operations after a crippling ransomware attack

🇩🇪 Black Semiconductor raises $273 million for graphene chips

🛰️ Apex raises $95 million for their off-the-shelf satellite buses

🌍 Kelvin is applying AI to home energy audits

Albedo

Albedo is a startup that will launch its first satellite in February 2025, thanks to SpaceX’s Falcon 9 rocket.

Their satellites will operate in very low Earth orbit (VLEO) and capture ultra-high-resolution images at 10 centimetres per pixel.

This is a big deal, as that kind of resolution used to be exclusive to American defence and intelligence companies.

It’s hoped that this can now be opened up to commercial businesses with much lower budgets, or maybe even the general public.

Albedo claims it can offer these images at very low prices, due to its satellite design being more efficient than their rivals.

They’ve already secured 7 customers - such as SkyFi and Open Grid Europe - who have reserved access to some of their satellite’s imagery. If you want to learn more about them, I’ve added a link below.

This Week’s Art

While the week has been dominated by Apple’s long list of announcements, it was a bit quieter compared to previous weeks. We’ve still covered quite a bit, including:

• Huge tariffs on Chinese EVs in the EU

• Elon Musk’s $56 billion pay deal

• Mistral AI’s new funding at a $6 billion valuation

• BeReal’s acquisition by the French company Voodoo

• How Generative AI is being used to train robots

• Apple’s long list of AI announcements and the wider impact

• Research into humanoid robots that can drive

• And how Albedo are trying to disrupt the satellite industry

Have a good week!

Liam

Share with Others

If you found something interesting in this week’s edition, feel free to share this newsletter with your colleagues.

About the Author

Liam McCormick is a Senior AI Engineer and works within Kainos' Innovation team. He identifies business value in emerging technologies, implements them, and then shares these insights with others.