Welcome to ICYMI – a weekly snapshot of European news stories that have given me pause for thought. ICYMI is a chance for you to go beyond the front-page headlines and find out what other stories may be worthy of your attention. This week:
- What are games teaching AI about strategy?
- How is hiring changing across Europe?
- Should we be afraid of Zombie Unicorns?
Earlier this month, I blogged about how Dario Amodei, the CEO and co-founder of Anthropic, raised a few eyebrows with his prediction that AI could surpass the greatest minds and Nobel prize-winners in most fields.
Fast forward a couple weeks, and now we’ve got an actual Nobel Prize winner – Demis Hassabis, CEO of DeepMind – offering a glimpse into the future.
Speaking with CNBC, Demis shared his prediction that artificial general intelligence (AGI) is just 5-10 years away. AGI could be capable of making complex decisions with real-world context, potentially matching or surpassing human-level decision making.
As part of developing AGI, Demis reveals that DeepMind is focusing on training agentic and multi-agent AI systems and is using the StarCraft video game to do so. Bizarrely, it turns out that intergalactic warfare could help AI models develop real-world reasoning.
But before the machines start winning Nobel prizes of their own, there’s still a lot of groundwork to be done first.
This week, researchers at Columbia Journalism Review raised fresh concerns about the reliability of ChatGPT, reporting that over 60% of its responses include false information.
While AI is improving every day, for now at least it’s a timely reminder that there’s still work to be done – not just in fixing inaccuracies and enhancing capabilities, but in building confidence around how and when these tools can be trusted.
That trust gap is showing up in other areas too. I was interested to read in the FT this week about a Harvard Business School study examining the use of AI in contact centres which found that 30-50% of customers are willing to wait significantly longer to speak with a human over an AI chatbot.
The same study also evaluated how people perceive the quality of these interactions. Participants were shown identical customer support messages but rated them more highly when told they were written by a human.
So even when the output is the same, perception matters – and winning hearts and minds may prove to be just as important as improving the technology itself.
What next for workers in the workplace of the future?
LinkedIn’s latest report on the most in-demand skills for 2025 shows that while AI skills like LLM development and AI literacy are skyrocketing in importance, it’s human skills such as communication, adaptability, and critical thinking that still top the list of what employers want.
This is mirrored in the Netherlands, where employers are placing greater value on softer skills such as problem-solving, collaboration, and adaptability over degrees and other qualifications. As AI takes over more routine tasks, the ability to grow, learn, and adapt are becoming some of the most valued strengths among workers.
But amid this rapid transformation, there’s also a growing human cost. A recent PwC report found that 40% of Gen Z workers are considering leaving their jobs due to mental health struggles – a sobering statistic that reflects deeper concerns about burnout, job insecurity, and feeling overwhelmed in a world that’s shifting beneath their feet.
In a labour market already grappling with skills shortages and demographic pressures, this is more than a wellbeing issue – it’s an economic one. If younger workers continue to disengage at this rate, it could lead to long-term productivity challenges and strain on public systems.
As more businesses embrace AI, supporting the wellbeing of staff and the development of new skills will be as crucial as investing in the tools themselves.
Zombie Unicorns and a reality check for French start-ups
In France, a wave of post-Series A bankruptcies is hitting the tech scene hard.
A new report from ScaleX Invest found that more than 10% of French startups that had successfully raised Series A funding are now facing serious financial trouble.
Even more sobering: in 2024, there were more insolvencies than new Series A rounds. That’s the first time the balance has tipped into the red.
What’s going wrong? The majority of these startups raised cash within the last three years, but struggled to turn funding into real, sustainable growth.
This must be a wake-up call – for founders and investors alike – to focus less on big raises and more on early profitability, solid fundamentals, and risk management. Something we explore at length in Tyto’s Scaling Without Borders podcast series.
Meanwhile, across Europe, some of the tech darlings of the last few years are starting to show cracks. Forto, Sorare, and Xentral – once hyped as unicorns – are now being labelled “Zombie Unicorns.” According to Manager Magazin, they’ve raised huge sums, but behind the scenes are burning cash, cutting back, and losing investor confidence.
It’s not just startups and scale-ups feeling the pressure. In the UK, over half of all business leaders are expecting a recession this year, according to BCG’s latest State of UK Business report.
Inflation and rising taxes top their list of worries, so I’ll be watching closely to see what the UK Government plans to do to address this when the Spring Budget is announced later this week
However, it’s not all doom and gloom. There was also some good news last week, with Oracle announcing plans to invest $5 billion in the UK over the next five years, as part of a package of initiatives design to improve cloud infrastructure and AI across the country.