In 2026, AI systems like Gemini 3.8 and OpenAI's Astra manage complex office workflows around the clock without breaks or office politics. As algorithms handle routine administration from spreadsheets to emails, the traditional desk job fades as a central anchor of life. We must urgently ask what direction we want to set for society with the time being freed up.
I first heard the term AI in 2018 – from engineers from TU Chemnitz whom I got to know because I was running a Business English workshop for an education provider that counts among the positive experiences of my career: indisoft. I did a lot of good things with indisoft, in Leipzig as well as in Dresden. Really pleasant work, I remember it very fondly – those were good projects. And one of these projects was that I gave these engineers the task of preparing mini-presentations on a particular topic in small teams – in English, of course – and giving them too. And two of these engineers picked the topic of AI – in 2018, so almost nine years ago. And what they reported was incredible: what I still remember is that they said there had already been cases in the US where the developers didn’t understand what the AI systems had built – that the human programmers could no longer follow the logic, the AI’s logic. That’s how far we already were back then.
Then it took a few more years, until the end of 2022, for the breakthrough to come, and suddenly OpenAI burst into the public eye out of nowhere and launched ChatGPT. And then – it must have been 2023 or 2024 – when I did my BAMF accreditation as a teacher with ChatGPT: that written work, that portfolio the BAMF requires, I wrote to a large extent with ChatGPT, and I’m not ashamed of it. I’m saying this quite explicitly here, because for me these BAMF requirements have nothing whatsoever to do with reality. And it’s the BAMF that needs a reality check, not me. And if they have these artificial requirements that have nothing to do with reality, then I’m not going to fritter away my time on them – of course I’m going to use a tool like ChatGPT. And I still worked my own ideas into it back then, but most of the wording came from ChatGPT. Interestingly, my portfolio got a very good assessment back then – so even then nobody noticed that it was mostly written by AI.
In any case, development has moved at breakneck speed since then. And I’ve just had another look at my entry number 296 from a few weeks ago: there I talked about AI systems like Gemini 3.8 or OpenAI Astra. But by now I already have access myself and am currently using Gemini 4 Argon – and that’s a huge leap forward. The difference in text production between Gemini 3.8 and Gemini 4 Argon is enormous, and this step, this development, happened within the last few weeks – so from the original date of entry 296 in September to now, the beginning of November, AI has already improved enormously.
And this development, this speed, this improvement will keep going. And that’s something we humans don’t understand, because we’re not used to exponential learning curves. We humans usually learn in a more linear way. Maybe there are a few breakthroughs now and then, and then things really take off – in language learning, for example, or other things – but then the growth stops again very quickly. And from a certain age the peak is behind us and we just don’t learn anymore; on the contrary, we decline. AI knows no decline. AI knows no standing still, AI knows no going back, no forgetting – and that’s something we haven’t grasped yet, because it’s completely new. We’ve never experienced anything like it.
What do you take away from this thought?
Thank you. Every answer fits – take it with you into your day. Would you like to share something yourself?
