You’re receiving this email because you subscribed to the newsletter. If you’d like to unsubscribe, you can manage your preferences here with one click (or use the button at the bottom of the email).

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!

1. Google plans an automated research lab in the UK

We start this week with DeepMind, which has announced new investment in the UK and plans for an “automated research lab”.

The new facility will be able to work seamlessly with Gemini, with robotics used to create and test hundreds of materials every day.

This will become a great testing ground for Google, which has been working towards this goal for years. I still have my doubts over LLMs and their genuine abilities to discover new materials, as they are only one part of the puzzle.

To get solid results and make these scientific discoveries, we need to wrap guardrails around these LLMs and integrate them with other AI systems - they can’t make these discoveries on their own.

UK scientists will also get priority access to DeepMind's "AI for Science" toolkit - including AlphaEvolve for designing advanced algorithms, AlphaGenome for DNA research, and WeatherNext for more accurate weather forecasting.

Google has also agreed to collaborate more closely with the UK AI Security Institute, as they will work together on testing advanced AI systems and reducing the harm to ordinary people.

Overall, it’s great news for the UK which has been crying out for investment. The technology still has quite a while to go before it becomes fully automated.

DeepMind is still making advances on controlling robots, so it will take several more years before that can be fully joined-up. But of all the major players - OpenAI, Microsoft, Meta, and Apple - DeepMind has been leading these research areas for years and that looks set to continue.

2. UK Government doubles down on quantum computing

Continuing with the UK, the government has announced that it will invest £670 million in promising startups - with much of it being concentrated on just a small handful of them.

It's a significant shift in strategy for a country that's historically spread its tech funding too thin, but this action was badly needed. Government officials were forced to act, following the recent acquisition of Oxford Ionics by an American rival.

Quantum leaders were worried about the lack of investment opportunities in the country, as they have seen rivals in other nations raise billions. That unease prompted them to look elsewhere, which has forced the UK government to respond and ensure that this critical research isn’t bought by others.

Previous governments have made the same mistake with AI, as ARM was bought by Japan’s SoftBank, so politicians were keen to avoid a repeat with quantum.

For some time now, the UK has been a quiet leader in quantum research - capturing 18% of global funding since 2020. But that advantage is at risk of slipping, with France investing €500 million in just five startups and the US government now prioritising the technology.

3. Helsing and Kongsberg to build Europe's first defence satellite network

Both companies plan to create a satellite constellation by 2029, allowing Europe to gather intelligence and mark new targets - completely independent of the US’ technology.

This is a direct response to lessons learned from the Ukraine war, where space-based reconnaissance has proven essential for modern warfare, and to changes in global politics.

Kongsberg’s expertise will be used to build the satellite hardware, while Helsing will add its own AI platform to monitor changes in the region and detect threats.

Helsing’s technology is already being used in battles over Ukraine, so it’s battle-tested and has been refined to spot Russian forces.

After years of relying on American technology companies and the US military, Europe is finally bolstering its own defence sector and will become more independent.

Germany has been the driving force behind this move and are keen for everything to be built within Europe - with Isar Aerospace being lined up as the preferred launch partner. If all goes to plan, these satellites will be launched from Norway's space complex in Andøya.

4. AI agents have arrived in the enterprise

Based on Anthropic’s State of AI Agents report, which surveyed over 500 US leaders, companies are now deploying the technology into production systems.

The headline numbers are pretty striking. 57% of organisations are now deploying agents for multi-step workflows, with 16% using them for complex tasks that span across multiple teams.

Almost 90% of companies are now using AI agents for software development, which isn’t surprising as Claude Code has taken the industry by storm this year.

Interestingly, 80% of leaders reported that their investments are already delivering measurable ROI. Quite frankly, I find this hard to believe.

I know that most companies are seeing productivity gains with AI coding assistants, but these agents do make mistakes and you need to carefully review what they generate.

If you don’t and become too relaxed as an organisation, you will find that overall productivity has actually fallen - as staff have to review the AI’s results and spend more time fixing them.

This is known as “AI slop” and is a real problem. I’m a bit skeptical of this ROI figure, so it would be good to see more data from independent organisations.

Regardless, companies have said that the highest-value use cases for agents are software development, generating reports, and automating internal processes.

The main barriers to adoption are the ones that we often hear about: difficult to integrate with other tools, poor data quality, and change management.

If you’re unsure about the technology, I’ve written this detailed guide that outlines how it works and how you can use it.

5. Tech companies setup a new body for AI agents

Continuing with the theme of AI agents, the Linux Foundation has launched a new initiative to standardise AI agents and how we connect with them.

With this new Agentic AI Foundation, tech companies - including OpenAI, Anthropic, Google, and AWS - will work together on new standards.

Anthropic has donated its MCP technology to the foundation, with OpenAI doing the same and giving it control of the AGENTS.md format.

It’s hoped that this will boost enterprise adoption of the technology, as it will prevent “closed wall” platforms and companies being locked into a specific ecosystem.

Autonomous drone swarms just took a major step forward

Well, it seems like warfare is about to change forever. Auterion has successfully co-ordinated a drone strike, with drones from multiple manufacturers being used.

All of the drones were connected with Auterion's swarm engine and were able to find the target, come up with an attack plan, and strike the target.

But the key thing here is the interoperability between manufacturers, as these drones were able to work together without bespoke integrations.

Following the war in Ukraine, Western militaries have invested significantly in drone technologies and being able to autonomously control them. This is another step towards an automated war-zone, as Russia uses drones that cannot be electronically jammed and continues to take territory across Ukraine.

It remains to be seen whether this is a polished demo from Auterion, which is based in the US, or a real breakthrough that Western allies can start to use.

But the direction of travel is clear - military planners are now betting on autonomous swarms that can scale across manufacturers and nations. The next challenge is securing them, so that they’re not used against us.

Google’s AI try-on feature now works with selfies

Google has upgraded its AI try-on feature, allowing users to generate a full-body digital avatar from just a selfie.

Previously, the tool required a full-body photo to work its magic, but the new approach uses Nano Banana to do this for you.

You can simply select your usual clothing size, then Google’s feature generates several images for you to choose from.

The feature was launched in the US this week, building on Google's original July release that allowed shoppers to virtually try on items.

Confusingly, Google has its own app - called Doppl - that does the same thing, but works similar to TikTok and shows you outfit ideas. Once you’ve created your virtual avatar, Google will suggest different outfits and show how they might look on you.

Once you’ve found one that you like, you can buy it in a few seconds. Essentially, this is an extension of Google’s advertising platform and allows companies to easily increase sales.

The downside is that Doppl is only available in the US, but it should be rolled out to other countries in the coming months.

🔒 Australia bans under 16 from social media

⚖️ Disney accuses Google of copyright infringement, signs licensing deal with OpenAI

🎧 Google Translate now lets you hear real-time translations in your headphones

🚗 Waymo's robotaxis did 14 million trips in 2025, over 3 times more than last year

📄 New RSL 1.0 standard supports pay-to-scrape for web publishers

⚡ Ford and Renault will build cheap EVs in Europe

🚀 SpaceX plans to IPO in 2026 at a $1.5 trillion valuation

🎬 Paramount launches a $108 billion hostile takeover bid for Warner Bros.

💡 AI reasoning models are use 30x more electricity than other models (Bloomberg)

🔬 Google launches an AI research agent

🤖 OpenAI releases GPT-5.2 and makes some performance gains

🌍 Runway releases its first world model

🔧 Google launches managed MCP servers and allow AI agents to access its tools

With Christmas coming up, I thought this startup was quite topical. Tavus is a company that allows families to video call Santa Claus and have conversations with him.

Of course, it’s an AI version of Father Christmas - but the quality is surprisingly good.

Since the website has access to your webcam, Santa can read your facial expressions, remember past conversations, and even search the web for gift ideas.

According to Tavus’ CEO, the company is already outpacing last year's millions of hits - with some users spending hours chatting with Santa Claus.

But that level of engagement with young children raises some uncomfortable questions. Kids can struggle to distinguish AI from reality, and extended chatbot interactions have already been linked to harm in adults.

The launch comes as the industry grapples with serious incidents, including Character.AI restricting under-18 access following reports connecting chatbots to teenage deaths.

Tavus says that safety measures are in place, with content filters used and controls that will automatically stop the conversation when needed.

If you want to try it out for yourself, I’ve included a link here.

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.

This will be my last post ahead of the Christmas break and I’ll have another round-up at the start of January. Wishing everyone a lovely holiday and a happy New Year!

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 Kainos' Innovation team. He identifies business value in emerging technologies, implements them, and then shares these insights with others.