One helpful thing is to decide what not to stay current on, because it is impossible to know everything. For instance, I stopped caring about new model updates from OpenAI, Google, Meta, etc. some time ago. Most of these new models do something complicated impressively well, but then it is also shown that they fail comedically in other, simpler tasks. However, I try to read the work that comes from top research institutes when they investigate how AI impacts people and societies. At the end of the day, whatever comes out of the “AI revolution” will have to go through a complex sociotechnical process, not just a technical one. Therefore, I also try to pay attention when someone who is not an AI expert (but whose perspectives on current developments I usually find informative) expresses an opinion on something AI related. I also find it helpful to remember that there is still a slower moving, non-generative, chatbot AI field that is also relevant to follow, sometimes maybe even more so than the fast-paced stuff.
I follow developments in a fairly informal way, for example, through TikTok, where videos sometimes appear showing how to use certain tools or tips and tricks for projects like the final degree project (TFG). These videos can be helpful for discovering practical uses of different applications or learning shortcuts that make tasks more efficient. However, beyond this casual exposure, I am not particularly well-informed about the topic. My knowledge is mostly limited to what I encounter in everyday social media content, and I haven’t actively sought out in-depth or structured information about these tools. While platforms like TikTok can provide interesting insights and quick tutorials, they don’t replace more thorough research or formal learning. As a result, I feel that my understanding is still quite basic, and I would need to explore more specialized resources or courses if I wanted to gain a deeper and more reliable knowledge of the subject.
The best way to get started in the world of AI is to combine openness with critical thinking: Trying out new tools and gaining your own experience is valuable – but it is crucial to always reflect on what a tool actually does and where its limits lie. At the same time, one should say goodbye to an unrealistic claim from the outset: no one can have complete overview knowledge in this area. The development is so rapid that even experts can hardly keep up. This is not a weakness, but a reality that everyone has to deal with equally. More important than knowing everything is therefore a basic attitude: stay curious, try things out, question sources and classify results. Anyone who approaches the topic in this way is well equipped – regardless of how quickly the technology develops.
Since my interest in the topic has continued since the free access to LLMs, it is not difficult for me to stay "up to date". However, this does not mean that I acquire every program that comes onto the market. It is important to know and be able to do one's field, to keep learning and to get involved in the matter. Ideally, several AI systems can be combined to generate outputs. The following also applies: Ask the AI if you don't know what to do. According to my observation, this option is underestimated and underused
In my case, that is part of my career. Software engineering, even before AI, has been in an ever moving field, and a constant curiosity and re-education is needed to keep up.
I follow specialized sources, scientific publications, and technology news, and participate in online courses and professional communities. I also experiment with new AI tools and exchange insights with other experts in the field.
I mix curated sources with hands-on testing—if I can’t try it myself or see real use cases, I don’t fully trust the hype.
