Articles in TLDR are turning negative about AI, but we don’t agree.

A list of negative articles in the latest TLDR newsletter has me wondering. Because it doesn't rhyme with our daily AI practice, or with the progress the EU is making.

AI has been hailed ever since the SF of the fifties. It is now come into reality, and is maturing fast.

Many areas of industry are being revolutionised by AI. And healthcare. Last year, DeepMind’s AlphaFold 3 was released, able to predict 3D structures such as DNA and RNA with unprecedented accuracy.

Critical knowledge gained, used by 1,000 research teams, enabling rapid progress in treatments for neurodegenerative diseases, antibiotics for superbugs, and personalised cancer therapies. Has the potential to cure 1 billion patients suffering from currently untreatable conditions.

But I look at the latest TLDR (TLDR Web Dev 2025-09-05), I see much gloom and doom about AI coming from the software industry. What’s that about?

“Where's the Shovelware? Why AI Coding Claims Don't Add Up”. Mike Judge on Substack is disappointed that promised speedups are not achieved.

“MIT says AI isn’t replacing you… it’s just wasting your boss’s money”. Interview Query Com says that companies are struggling to get ROI out of AI.

“Vibe Coding Our Way to Disaster”. According to Jake Nations on Anthropod Software we are losing the art of 'thinking deeply' by using vibe coding.

They do pivot a bit at the end. But it’s obvious that they are gloomy about AI.

There's adoption issues. There’s deep fakes, but fake news been around since the printing press.

In the meantime, here in the EU, good things are happening.

Yesterday, it was announced that the supercomputer Jupiter went into production. It's Europe’s first exascale supercomputer, at 1 exaflop (1 quintillion calculations p/s).

To be honest, to close the gap with the US, we need 3–5 exascale systems by 2030, increase data centre capacity by 50%, invest €15B to make them green and energy-efficient, double EU private investment to €20B/y by 2030, up the number of EU startups by 150%, train 50,000+ new AI specialists (or attract those fleeing the US). And more.

Jupiter is a symbol of EU AI sovereignty. With exascale + quantum-ready compute, investments in talent and data, regulatory clarity (AI Act), industrial collaboration, the EU can close the gap by 2030 and overtake the US in sustainable, human-centric AI by 2035.

That’s EU’s ten year plan. China recently published theirs: to become the dominant AI power by 2035. The US? Well, they don’t make plans anymore.

For us as a software company, these last couple of months has been an exciting time. Using AI and vibe coding, we see time spent coding diminishing day by day. We add features to assignments just for the fun of it. When we worry about effort, we worry how we are going to explain to the client that they got so much more for their money.

And we learn to take articles in TLDR seriously, but not too much so. Because we learn to look to ourselves, at how the EU is doing, how much AI can help us.