Weeknotes 276 – Is the end of prompt engineering near?

Hi, y’all!

Welcome to the newsletter, especially the new subscribers. There is a lot of interesting in the news going beyond the current platforms and new introductions in AI capabilities, making movies that also populate mainstream media. At the same time, Google is upgrading Gemini with some interesting features that might change prompt engineering, which definitely triggered a thought!

Before I continue, let me know if you want to dive deeper into these topics or create specific (near) future studies. Check out the Target_is_New website to find out what is possible.

Triggered thoughts

Apparently oio.studio did work Google’s Gemini new interface. That made me curious as I know them for their interesting thought experiments, as well as real experiments with more-than-human futures, partnerships with autonomous creatures, and creative AI overall.

The Gemini explanation shows how they create a reasoning interface that makes a debriefing of the intention of the human input. It makes me think about how we get more human interactions with machines, as we don’t need to adapt to machine-like thinking anymore to get the best. We got used to adapting our thinking a bit to the machine way of thinking, like with search engines. The first step was the translations from natural language to machine-understandable queries in the chat interfaces. Gemini seems to go further by adding an extra layer of ‘understanding’. So it is not only the interface like with chat interfaces. It is also the internal reasoning design that is made differently. In the demo, you see how, under water, Gemini is trying to make sense of what type of question it really is and is adapting the interface to that type.

Would this mean that the profession of prompt engineering is ending soon? For now, the example shown focuses mainly on presenting the search results. It is unclear if it is adjusting the findings, too, and if there is real understanding (see below). As a new approach of shaping the interface, it is following what Perplexity.ai is also trying (check out the interview with the founder at last week’s Hard Fork); finding a new form for building relations with our AI partners.

Check the newsletter for all the notions from last week’s news.