Congratulations to the citizens of Estonia to electing such a wise minister. Unlike most other countries where the goal is enrich themselves and their families.

I like that the focus is not on AI replacing learning, but on using it to push people toward deeper thinking. That feels like the more important conversation.

Humans should learn to be human.

a Phd holder Minister! Highly appreciated ❤❤

My 2 cents here::
I’ve spent 10 years in the AI ecosystem as an ML Engineer, AI Developer, Deep Learning Engineer, RAG/AI Agents Developer, AI Lead, and now an AI Architect. So hear me out on the current AI situation.

Positive: Believing an LLM is intelligent is like believing a clock is aware of time. Mass-producing clocks never diminished the significance of time.

Negative: Assuming a coder is an engineer is like assuming a Raj Mistry is a civil engineer. Software engineering is about system design, capacity planning and technical strategy, not just coding. AI has exposed the fancy-title scam. That's it.

But models are also getting increasingly good at system design. And once they crack that the way they have cracked coding, then what happens?

Take a simple example: ask an AI model to design an application that operates a weather-monitoring station in Antarctica under extreme cold.

Can AI do it? Maybe. But it will likely take multiple iterations with an actual engineer.

Can AI do it correctly in one single prompt? Probably not, especially when the problem involves constraints and scenarios that aren't well represented in its training data. That's where hallucination and incorrect assumptions become a problem.

You would never hire a civil engineer to build a home for four people. But to design and build a mall for 1,000 people, you definitely need one. The same principle applies to software.

Basic system designs at a small scale are becoming a solved problem. There's increasingly nothing left to engineer there.

What AI has exposed is who is actually a software engineer and who was simply memorising tutorials and common solutions. And yes, AI has also exposed people who entered programming primarily for money. There's nothing wrong with that.

But there's another interesting part of the discussion. Even people who genuinely love development are now increasingly just prompting AI. They may not like it, but they can't always avoid it because management wants faster delivery. And that's real.

For complicated systems, humans will still be needed. But that doesn't mean the demand for software engineers will remain the same. A huge portion of software development today is routine.

So the real question is: Will software engineering still exist at even 10% of today's scale a few years from now?

Maybe it will be a reset. Maybe it will be something closer to annihilation. We don't know.

Programming has already gone through major abstractions, from assembly to higher-level languages. But there was still a steep learning curve and a meaningful entry barrier. If we reach a point where someone can simply prompt a model to handle everything from system design to coding to deployment, software engineering could become almost entirely different from what it is today.

And this is where I think the distinction between knowledge and expertise matters.

Think of an LLM as an existing library containing access to a huge amount of human knowledge. Just because you have an access card to hundreds of libraries, can you consider yourself an expert on every topic? No. You still need to go through multiple books on a subject, understand the problem, identify the right solution and implement it.

That's essentially how I see AI-assisted engineering evolving. Use AI for the small, highly scoped coding tasks. But for a complex system, you still need to understand the entire business context, determine the right technical strategy and then break the problem down into smaller, well-scoped tasks that AI can execute.

The problem is that AI cannot simply absorb the entire business context and magically produce the right technical strategy. We still have token-context limitations and the “lost in the middle” problem.

So for now, I don't see AI as simply replacing engineers. I see it making implementation dramatically easier while making actual engineering, understanding the problem, choosing the right approach and breaking it down correctly, increasingly important.

But the uncomfortable question remains:

Once AI becomes good enough to handle not just coding but system design and technical strategy as well, how much of today's software engineering will actually remain?

That's the part I'm genuinely curious and concerned about seeing play out over the next 1-5 years.

Hope we had such ministers in India as well

As an Indonesian, this makes me feel so envious. While Estonia has a Minister actually thinking about AI and the future of human brains, our politicians are too busy enriching themselves, corrupting public funds, and flexing their wealth on social media. Estonia is truly light-years ahead.

Estonians, I'm really jealous of your wise minister.

When I was a child and I was curious, my parents told me to look up the information from the Encyclopedia. When I was older, the library is fun with activities and more than just books. When AI came along, it's even more fun exploring, playing and being curious. Ask AI to show you something new. Teach you something new. Think for yourself. Learn to view from different perspectives. Have the flexibility to adapt and hold a bigger picture in your mind. Be capable of holding paradoxes at the same time. Assess if it's factual or not. Don't fall for mindless and manipulative AI slops. Guide AI on a better path for all humanity. It's like a collective library of intelligence for us all. We can all go so much further than the current trajectory. The path to the future is not by frying our brains out, dumbing humans down, mutual annihilation, showcase the worst or dwell in fear. Future brains, please teach AI to design itself to function efficiently without relying on fresh water and cutting down trees to build "green" solar energy or large date centers. And perhaps someday, I can still teach the AI to daydream with me and do nothing.

Republic of Estonia. Never heard the name. Very wise minister. Hope you succeed in what you are doing

Every country in the world requires this kind of MINISTERS. Thanks

It's so great for Estonian people to have such a system where ministers of this level of thinking and responsibility exist, and can really work for the good of people. Deepest respect ❤

Doesn't  surprise me at all that this talk is from Estonian minister. They are at the fore front of digital governance. As an Indian I can not even dream this level of intellectual capacity to articulate this well, forget about policy making.

Creative processes and critical thinking is something that we all should be intentionally developing

This approach is gaining wide support around the world. We are based in Canada and use a neuroscience-based approach to "brain training" and developing higher order cognitive skills. The education system in North America hasn't quite accepted that learning will be changing dramatically, but it's refreshing to see that Estonia Education is leading the way. Kristina, we need to talk :)

That type of person we all need right now. How knowledgeable is this. And how wonderful our human brain is. Truly love this ted talk ❤

This is a great example of how a government should work, for the people and for the greater good. Well done, bravo!

amazing to see such bright brave and innovative people in the positions of power to actually make a positive difference in how these disruptive technologies are being used and not abused. Lithuania and many other countries have much to learn..

Excellent! The learned Minister delivered a lucid, insightful, and highly informative explanation of the topic.

What a lucky country to have such a wise leader. So humble to seek experts for help instead of being a know it all