You’re lost when it comes to AI. That’s normal.

Artificial intelligence takes up an enormous amount of space in the tech news feed. Announcements arrive every day, at a pace that is hard to keep up with. There is a simple explanation for feeling overwhelmed. Even after following the subject for years, I can clearly see that the last two or three years have turned the volume of information into an almost unmanageable mass.

AI announcements arrive daily

OpenAI illustrates this acceleration well with its 12 days of innovations, a sequence built around one new development per day. This pace also fuels a battle for publicity between tech giants, including Google, Meta and OpenAI. More recently, Chinese companies such as Tencent and ByteDance have gained visibility in this race. Tencent, for example, unveiled a video generation model presented as more capable than OpenAI’s Sora. Information piles up quickly. The time available to verify, compare and test things properly remains much more limited.

The AI news feed overflows for two reasons

Two factors explain much of this frenzy.

The first is the explosion in the number of platforms. New tools appear constantly, each with its own announcements and demonstrations. Updates add even more noise. Discovering new players in AI is becoming very easy. Identifying those that actually matter takes much more time.

The second relates to the visual impact of generative AI. The results are immediately visible. A generated image or a video produced by a model can be understood in a few seconds. An interface that responds in natural language also gives a very clear impression of progress. As early as 2011, the rise of computer vision had attracted media attention thanks to its visual applications. Today, generative AI amplifies this mechanism with systems capable of producing images, videos or music from a prompt. This visibility directly feeds publicity around these technologies.

Filtering becomes a survival strategy

Faced with this overabundance, keeping your head above water primarily requires a method for filtering information.

• Accept limits. Nobody can follow everything. This is also true for AI professionals in a field that changes so quickly. (we’ll discuss this in a planned article)