Welcome to this edition of Loop!

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

‏‏‎ ‎ HIGHLIGHTS ‏‏‎ ‎

Let's jump in!

Most platform evaluations start with a demo request and end three weeks later in a conference room. This one takes 15 minutes and puts you directly inside Gladly's interface — navigating it on your own terms.

See how AI surfaces real-time customer context before a conversation starts. Watch how a single conversation thread pulls in purchase history, channel history, and account details without a handoff.

No installation. No commitment. Start the interactive demo and see the platform for yourself.

See it yourself

1. Apple sues OpenAI over trade secrets

We start this week with Apple, which has sued OpenAI for stealing trade secrets and using them to build its own consumer hardware.

The suit names OpenAI's hardware chief Tang Tan, who worked as an Apple executive for 24 years, and claims that he emailed himself information about Apple's suppliers.

It's also alleged that Tan asked Apple employees to bring "actual parts" from Apple to their job interviews, so that they could be shown to the OpenAI team.

Another OpenAI employee is alleged to have downloaded dozens of confidential hardware files from Apple, after they gained access to their former colleague's laptop.

That same employee then went on to coach other people on how they could avoid being caught by Apple's security team.

Overall, it's a remarkable breakdown in the relationship between the two companies. Just two years ago, Sam Altman announced that ChatGPT would be integrated into Siri.

Since then, OpenAI has acquired Jony Ive's design company and Apple has decided to partner with Google instead - with their technology being more tightly integrated within iOS.

2. Instagram and Facebook may need a redesign after EU ruling

Over in the European Union, politicians are concerned that the "addictive design" of Facebook and Instagram may have broken the Digital Services Act.

This focuses on specific features like autoplay, infinite scroll, and personalised recommendations, which are pushing users into compulsive scrolling and unhealthy habits.

Meta disputes the EU's findings, pointing to a feature called Teen Accounts that allows parents to restrict screen time and block children from accessing it at night.

The trouble is that these safety tools don't seem to hold up in practice, as a recent study found that two thirds of Instagram's settings were either broken or too hard to find.

If these preliminary findings are confirmed by the EU, Meta could be fined up to 6% of its global revenue - which would be over $12 billion.

The commission wants Meta to disable autoplay and infinite scroll entirely, which would be a fundamental change to how these apps work.

It remains to be seen whether the EU follows through and takes action against Meta, but given that several countries are already moving to block under-16s from social media - it's safe to say that this is only heading in one direction.

3. Microsoft tries to cut costs by using its own AI models

Over the past few months, we’ve seen companies use AI models more often and begin to integrate the technology into their workflows. This has sent both demand and costs soaring, which companies are now trying to get under control.

Microsoft isn't immune to these pressures either, as some Excel and Word prompts are now being routed to their own models - rather than paying OpenAI and Anthropic for every request.

It's a notable shift for the company, as Microsoft previously advertised the fact that large parts of Office 365 were powered by those third-party models. This is part of a wider cost-cutting trend that we've seen in recent months, as Amazon, Uber, Meta, and Accenture are also trying to reduce their spending on AI.

While some users might not like it, this is a pretty sensible move from Microsoft as not all prompts in Word will need to use a frontier model. If the task is simple enough, you can simply assign it to a smaller model and save significant amounts of money.

For companies that are integrating AI models into their products, this is worth considering - as the cost of compute is not going to go down anytime soon. Over the last few years, there has been a huge jump in intelligence with these AI models - so smaller models are the best choice for most tasks.

4. OpenAI launches GPT-5.6 and ChatGPT Work

GPT-5.6 has finally been rolled out to the wider public, after it was blocked for weeks by the US government.

Interestingly, OpenAI is marketing it as a cheaper alternative to rival models, which reflects how much pressure customers are putting on labs over rising AI costs.

While it can feel quite overwhelming with a new model being released every few weeks, GPT-5.6 does seem to be a noticeable step forward in intelligence.

It's especially good at programming and seems to be on-par with Anthropic's Fable model, but at a lower price. Following Anthropic's decision to remove Fable from their subscription, I'm quite tempted by GPT-5.6 and will test it more thoroughly in the coming weeks.

Alongside the new model, OpenAI has announced ChatGPT Work, which is an agent that combines ChatGPT and Codex. This allows non-technical people to more easily use the platform and adopt it into their workflows - similar to Anthropic's Cowork.

You can use ChatGPT Work to quickly create documents, spreadsheets, presentations, and then share them with others in your team. It can also connect with the different tools you use, like Slack, Gmail, and Google Drive.

This means that you don't have to constantly paste info into your chatbot and then copy the response again. Instead, ChatGPT Work can automatically fetch this info and keep going until it has a finished document.

5. Global PC sales fall due to demand for AI data centres

With tech companies investing billions to build new data centres across the globe, there has been a sharp rise in demand for computer components - from chips to memory.

But this has only started to bite in the last 12 months, as RAM prices are climbing at an alarming rate and have forced computer manufacturers to raise their prices for consumers.

As a result, global PC shipments have fallen by nearly 5% to 68.2 million units, which is the first decline in two years. It's believed there will be an even steeper decline later this year, as current inventory begins to run out and further price rises hit.

Surprisingly, PC makers are still making more money despite selling fewer machines - with consumers absorbing the cost instead.

Industry experts expect that the memory shortage will only get worse over time and won't ease until 2028 at the earliest, which is bad news for the average consumer.

If you've been thinking about buying a new laptop, I'd buy it sooner rather than later, as prices won't be coming down anytime soon.

Notion launches Ship OS to manage your AI agents

Notion is launching Ship OS, a pre-built system that allows you to run your entire product development cycle in one platform - with agents handling most of the routine work.

Their platform is able to analyse messages from your Slack channel, draft PRDs, generate engineering tasks, and then assign it to a coding agent that completes the work for you. Once that's done, your team can review the work and decide whether further changes are needed.

Notion clearly wants to become the orchestration layer for agents, rather than having them scattered across separate tools like Claude, Cursor, and Codex. Whether that happens will come down to the engineers, as they're the ones who will adopt these tools and shape how they're used in practice.

Reviewing an agent's pull request is often harder than writing the code yourself, as you're blind to the changes that are being made and can easily default to accepting what was done - which has its own risks.

Notion does seem to be aware of this and has just launched HTML blocks, which allow the agent to visually show what was changed, but it'll be interesting to see how this turns out in practice.

📈 Samsung overtakes Nvidia to become the world's most profitable company

💸 SK Hynix raises $26.5 billion, the biggest foreign IPO in US history

🌍 Microsoft's carbon emissions went up by 25% last year

🍎 Broadcom will make custom chips for Apple

🌐 OpenAI shuts down Atlas, its AI browser

📊 Anthropic, OpenAI, and SpaceX are bigger than the last 25 years of tech exits

🗣️ OpenAI releases new voice models, which sound more natural

⚡ ZML releases a free tool that will speed up inference for AI chips

💰 AI chip maker SambaNova raises $1 billion at $11 billion valuation

🪖 US autonomous ground vehicles are fighting in Ukraine

This startup is based in Shenzhen and was founded in 2023 by former Apple engineers, including their CEO Will Wang, who worked on the Apple Watch and iPhone.

The company is developing its own smart glasses, but they've taken a different approach to Meta and Snap by skipping the camera entirely.

Instead, their glasses use a heads-up display that's built into the frames and shows information in your line of sight. The device is controlled by a companion ring, which you can use to tap and swipe.

The company has just raised $150 million in a round that was led by Meituan and Tencent, which values them at $1 billion.

Their camera-free design is central to their privacy philosophy, with voice features being used to transcribe audio into text - rather than storing the recordings - and all of this data is encrypted.

Their most popular feature is Conversate, which can understand your conversations in real-time and explain unfamiliar jargon - with follow-up questions suggested by the AI model.

Of course, removing the camera doesn't remove every concern - as these glasses are still listening to everything around you.

Imagine wearing them in a confidential meeting. Even with guarantees that the audio is never stored, people speak differently around a device that's transcribing them, and that could really stifle open debate.

Regardless, it's good to see a different approach being taken. As Meta and Snap continue to push hard on camera-equipped glasses, it'll be interesting to see whether the privacy-first bet pays off.

This Week’s Art

We’ve covered quite a bit this week, including:

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

Or if you’re interested in chatting with me about the above, simply reply to this email and I’ll get back to you.

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.

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