The train booked, the operator believes that its work has been completed. But I need to stay somewhere, I remind it – can it book a hotel? It asked for more details, and I purposely vaguely stated that it should be comfortable and conveniently located. Comparing hotels is perhaps my least favorite aspect of travel plans, so I was happy to scroll through booking.com. When I see it sets the wrong date, I can’t jump in, but it can correct this itself. It took a while to investigate the Ibis listing, but ended up choosing a three-star hotel called Martin’s Brugge, and I noticed the user’s location was excellent.
Now what’s left is the itinerary. Here, the operator seems to have lost motivation. It provides a perfunctory day-long schedule that seems to be mostly crib-cut from vegetarian travel blogs. On day 2, this indicates that I “visit any remaining attractions or museums”. Wow, thanks for the tip.
The day of the trip arrived and when I dragged out of bed at 4:30 AM, I remember why I usually avoided early departures. Despite this, I went to Brussels without any problem. My tickets were available to continue traveling, but I realized I didn’t know where to go. I fired the operator on my phone and asked which platform the next Bruges train departs from. It searches for Belgian railway schedules. After a few minutes, it was still searching. I looked up and looked at the details on the station display. I arrived at the platform before the operator figured it out.
Bruges was a great time. Given the operator’s lack of travel, I branched out. I realized that this kind of research task is perfect for large language models – it does not require proxy functionality. The operator’s OpenAi sibling Chatgpt offers me a more thorough plan for drawing events by hour, not only where to eat, but what to order (at De Halve Mann Brewery’s Flemish Stew). I also tried Google’s Gemini and Human Claude, whose plans were similar: walk to the Market Square; see the tower of the bell tower; visit the Cathedral of the Holy Blood. Bruges is a small city and I can’t help but ask if this is just a standard travel route or if AI models all get information from the same sources.
Various travel-specific AI tools are trying to break through this versatility. I briefly tried MindTrip, which provides maps with written itineraries, which provide personalized quiz-based advice and includes collaborative features for shared trips. CEO Andy Moss said it expands the wide range of LLM capabilities by leveraging a travel “knowledge base” that includes things like weather data and real-time availability.
Courtesy of Victoria Turks
