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!

Meet Accio Work, the agentic workspace designed to run your business operations end to end. From sourcing products and negotiating with suppliers to managing your store and launching marketing campaigns, Accio Work handles the execution so you don’t have to.

Powered by verified capabilities and deep integrations with business tools, it doesn’t just generate ideas, it takes action. Backed by Alibaba.com’s global supplier network and over 1B products, it seamlessly connects strategy to execution.

Stay in control while everything runs on autopilot.

Try One Week Free!

1. Finance leaders hold a crisis meeting over Claude Mythos

We start this week with Anthropic's Claude Mythos model, which has alarmed finance ministers, central bankers, and some of the biggest names in global finance.

Mythos was raised at an IMF meeting in Washington this week. The Bank of England, Barclays, and the US Treasury have recently been given early access to the model, so that they can understand how it works and test their own systems for vulnerabilities.

Officials are worried because modern finance is deeply interconnected. Banks, trading platforms, and payment systems all run on shared infrastructure - and a vulnerability in one part of that chain could ripple across the entire system.

The Bank of England's governor was pretty direct about it - if AI makes it easier to find weaknesses in core IT systems, bad actors will try to exploit them. Plus, financial systems are some of the highest-value targets in the world.

Fintech apps, payment processors, and the third-party software that many banks depend on are all potential targets. If the model is able to find a vulnerability in any one of those, it could trigger a chain reaction and affect the wider financial system.

Of course, this kind of technology won't stay exclusive for long and other companies will eventually build a similar AI model. Once attackers have access to these advanced tools, the financial sector needs to be ready and have their own security issues fixed.

2. Science Corp prepares to place its first sensor in a human brain

Science Corp is a leading brain-computer interface startup, which is now preparing to place a sensor on top of a human brain.

Unlike Neuralink, which uses metal electrodes that can damage brain tissue over time, Science is betting on biology. Instead, the company will grow neurons in a lab and then embed them into a sensor that connects with the brain naturally.

Since the device is about the size of a pea and sits on top of the brain rather than inside it, Science says it doesn't need FDA approval.

If it works, this kind of technology could eventually help treat conditions like Parkinson's disease, where current treatments can manage symptoms but can't stop the disease from progressing.

You might remember Science Corp from a few weeks ago, when I covered their eye implant that's helping blind patients read again. That's their most advanced product right now and it could be available in Europe as early as this summer.

The company has raised another $230 million last month and is now valued at $1.5 billion, which gives them plenty of money to fund these research projects.

The biohybrid approach is genuinely interesting and early tests in mice have been promising. If the science holds up in humans, it could solve the durability problem that every other BCI company is facing.

3. Robots have captured an enemy position, for first time in history

Like something from the Terminator movies, Ukrainian forces have captured a Russian army position using only robots and drones - with no infantry involved. This is the first time this has ever happened in history.

President Zelenskyy has confirmed that the operation was carried out entirely by unmanned ground systems and aerial drones, which forced the Russian soldiers to surrender.

Ukraine has been rapidly scaling up its use of ground robots on the front line, which are currently controlled by humans from a safe distance. They're now being used to attack Russian positions and pull wounded troops out of areas that are too dangerous for human evacuation teams.

It follows a similar advance back in January, when a Ukrainian ground robot used AI to autonomously detect and track targets. The robot was able to capture three Russian soldiers as prisoners of war entirely on its own.

When you take a step back here and look at the broader conflict, there are some truly staggering numbers. In the first three months of this year alone, ground robots have been used in over 22,000 missions on the front line.

This is a significant moment for how wars are fought. Ukraine has essentially become the world's testing ground for autonomous military technology, with both sides accelerating their use of robotics and AI as they look for any edge they can get.

Western governments are watching closely too. The lessons being learned on the front line in Ukraine are shaping how NATO countries think about the future of their own armed forces - and it seems to be moving a lot faster than anyone expected.

4. LinkedIn data shows that AI isn't to blame for hiring decline... yet

One of the biggest questions in tech right now is whether AI is starting to replace jobs. But according to LinkedIn, the answer is no - at least not yet.

The company has analysed data from over a billion people and haven't seen the kind of impact you'd expect, if AI was starting to displace workers.

Hiring in general has dropped by around 20% since 2022, but LinkedIn says that it's more closely tied to rising interest rates than anything to do with AI. This is something that I've flagged several times before, as companies overhired during the covid years and now need to scale back again.

Surprisingly, graduates don't seem to be any worse off either - LinkedIn's data shows they're being affected at roughly the same rate as everyone else.

While AI isn't displacing jobs at the rate that some people expected, it is changing them. LinkedIn predicts that the skills needed for the average job will change significantly by 2030.

The expectation now from business leaders is that people should do more work, at a faster pace, and at a lower cost - which is putting real pressure on how people approach their work.

The risk is that companies will push everyone to use AI for everything, when really they should be focusing on using it for the right tasks.

5. Robots outrun humans in Beijing half-marathon

For the first time, robots have beaten humans at running a half-marathon. Over 100 humanoid robots raced alongside professional athletes in Beijing on Sunday, and several of them were faster.

The winning robot finished in 50 minutes and 26 seconds, which is several minutes faster than the human world record. It's a huge improvement on last year, which was a bit of a disaster - most of the robots couldn't even finish, and many struggled to get off the starting line.

That said, the winning robot did crash into a railing just metres from the finish and had to be helped back onto its feet by humans. Although to be fair, I'd probably crash into a railing at the end of a half-marathon too.

It's clear that humanoid robots are a strategic priority for China, and the pace of development over the last year shows it's paying off. The government has been investing heavily in local companies and the technology was even given a prime spot on China's most-popular TV show, earlier this year.

It's fascinating to see the technology improve so quickly and to see China running these kinds of events. You have to admire the ambition here, as no other country is running a half-marathon with robots.

Anthropic’s new design tool is impressive

Anthropic has launched Claude Design, a new tool that allows users to create visuals with their AI model. For example, you can use it to design a prototype, presentation deck, and a one-pager.

The idea is pretty simple - you simply describe what you want, Claude generates an initial version, and then you refine it from there with follow-up messages.

It's aimed at people like founders and product managers, who need to go from an idea to something visual quickly, but don't have a design background - so it's not really trying to replace Canva.

When you get started, the tool will ask for your company's website and then scans it to automatically work out the different colours, fonts, and UI components that are used. It will then use these to create your own design system, which you can ask questions about and improve over time.

But if you don't have a company brand yet, you can ask Claude to help you create a new one from scratch.

I've been trying out the tool for a few days, and I have to say I'm really impressed with how it works. It's a great way for non-technical people to see designs change in real time and build on them.

In some ways, I think it's more useful for founders than Claude Code - especially if you need to explore different ideas visually before handing them over to Claude Code to actually build.

However, I did hit the weekly usage limit pretty quickly, even on the Max plan which has higher rate limits. It's worth noting though that Claude Design doesn't eat into your usage for other Anthropic tools, so you can keep using them as much as you normally do.

Claude Code has already become the go-to tool for a lot of software engineers. Now with Claude Design, Anthropic is building for designers too. But the challenge is that these are two very different groups of people who think and work in completely different ways.

Designers focus on how something looks and feels, while developers care about how it works under the hood. Getting both sides to collaborate smoothly has always been one of the hardest parts of building a product.

If Anthropic can bring both into the same ecosystem, it could change how teams work together - from the initial idea all the way through to the finished product.

🕶️ Apple is testing four designs for upcoming smart glasses

📊 Stanford report highlights a growing disconnect between AI insiders and everyone else

⏰ Claude Code introduces scheduled tasks, called "routines"

💻 OpenAI's Codex can now operate desktop apps on your computer

💰 Cursor in talks to raise $2 billion at $50 billion valuation, as enterprise growth surges

⚡ EVs are more affordable than gas (petrol) cars in the UK now

🤖 Physical Intelligence says its new robot can figure out tasks it was never taught

🔍 Boston Dynamics' robot can now read gauges and thermometers, thanks to Google's AI

🛰️ Amazon buys Globalstar for $11.57 billion to flesh out its satellite business

☢️ Inertia partners with the National Ignition Facility to commercialise fusion power

🚕 Uber and Nuro are testing a premium robotaxi service in San Francisco

🚀 Kepler launches the world's largest orbital compute cluster into space

This LA startup is building the "fastest unmanned aircraft" in the world and plans to reach Mach 5 - roughly five times the speed of sound.

Hermeus was founded by AJ Piplica, who previously co-founded Neuralink with Elon Musk, and has now raised $350 million in a round that values the company at $1 billion.

The company originally planned to develop its own engine, but eventually decided to stop and partner with Pratt & Whitney - one of the world's largest aerospace engine manufacturers - to modify their existing F100 engine for hypersonic flight instead.

That turned out to be a pretty smart move. Rather than spending years on engine development, Hermeus could focus on building and testing aircraft while picking up contracts with the US Department of Defense along the way.

Last month, Hermeus flew a demonstrator that was the size of an F-16, and they're now working on a supersonic version, with a third aircraft already in development.

The company is taking a SpaceX-style approach to aviation - build, test, fail, learn, repeat - which is pretty unusual in an industry where new aircraft programmes typically take 20 to 25 years.

Over $9 billion was invested in defence tech globally last year and Hermeus is well positioned to benefit from that trend. The company now has nearly 300 employees and the runway to keep iterating quickly.

If Hermeus can deliver on Mach 5, the military applications are pretty significant. You're talking about aircraft that could intercept hypersonic missiles, carry out rapid reconnaissance, or reach a target on the other side of the world before the enemy even knows it's coming.

It's great to see a defence startup moving at this pace. It'll be interesting to see whether Hermeus can actually reach Mach 5, but with two successful test flights already behind them, they're off to a strong start.

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