Design & Plans for Curiosio v18
by Vas Mylko
We released the most beautiful version of Curiosio exactly before the Ruzzian missile strikes all over Ukraine and the large-scale invasion of Ukraine. That version (v17) of the product received good feedback from travelers and supertravelers. Here is a making story if you have not seen it yet.
Excited and pleased with the success we planned for the several next versions to build up on top of the momentum. But here we are — it’s 107 days since then, 105 days of war — and we are only sharing the plans for the next v18. It’s a big war in Ukraine.
In this post, I will introduce [again] what Curiosio is and isn’t, how to use the current product for your benefit, tricks & hacks on how to get more from it. Then, will outline the scope of v18 and describe how we are designing it.
We are building the smartest search & answering engine for travelers. It is called Curiosio because it is embracing the mental process of the curious travelers. Follow your curiosity. Always follow your curiosity to see all places of the world [on the planet Earth] during your lifetime.
Curiosio is focused on road trips. You could travel in your car or fly anywhere by plane, rent a car and drive or ride there. Cars give you the ultimate flexibility, independence, personalization, privacy, and style. Here is our vision of modern and future car travel.
Economy class on the planes is becoming more and more frugal. Road cars are becoming more and more comfortable. The future of travel is inside the gadget.
Different travelers have different resources or capabilities of how much time or money they can allocate for travel. Curiosio is building the best possible experience in any requirements and constraints. You can create an interesting trip plan from scratch by the perimeter of interest, points/places, days, and budget. You can take other travelers’ trips and augment them by points, time, or money to your liking. You are asking Curiosio what you want and Curiosio is teaching you what is possible.
Curiosio has a search interface, to make searching & researching as simple as googling. We expected Google in travel and kept an eye open on the Google Travel portal. Back in 2018, they experimented with multi-leg air trips and multi-modal trips. There were usually multi-point journeys labeled with “Based on actual visits”. There were short destination trips, often labeled with “Written by Google”. We liked the idea of the feedback loop between the previous travelers and the current travelers planning new trips.
But Google Travel did not plan anything for you. It only connected your planning to the inventory for booking — flights, hotels, places. Back then we saw the opportunity that Curiosio could close this gap. Hence we started up.
Planning a multi-point journey within desires and constraints is hard. It is NP-hard. Only temporal and financial constraints are making the mathematical problem NP-complete. Curiosio allows adding geographical requirements such as geo perimeter, waypoints, travel-through points and places on top of time and money requirements. We cracked this math problem with our Evolutionary AI engine Ingeenee.
Scope of v18
Humans with AI are smarter than humans alone and smarter than AI alone. Ever watched Terminator 2: Judgment Day? The deep idea is that humans with T-800 outsmarted T-1000 alone. We are continuing the analogy — travelers with AI are getting more experience than travelers alone, or travelers fully relying on machine/algorithm.
We are going to introduce UX to kick off feedback loops. Visually it will be a second carousel with trip plans. Those trip plans will be selected from the ones liked by other travelers. You as a traveler and user also could like the trips created by yourself and created by others.
The feedback loops are happening between travelers, between travelers and AI, and between AI and AI. Travelers will not know whether the trip plans were created by other travelers, crawled on the web, or created by the AI algorithm. It will be the beginning of the data flywheel.
Conceptually it’s two feedback loops: User Loop with AI inside and AI Loop with Users inside.
Users create trip plans with unique requirements from scratch with Curiosio. Users augment and customize others’ trips with Curiosio (it’s a Supertrip function). Curiosio allows anonymous users to plan road trips. Such trip plans will be available for public view. Other users could see them in the second carousel, like them, promote via liking. Curiosio decides which of them to show and when. Signed-in users could keep their search results and trips private [by default].
Users see interesting trip plans in the second [new] carousel. Users could like them, thus increasing the chances those plans will be shown to more people. Users don’t know whether other users made these trip plans for themselves on the Curiosio platform, or they were crawled and captured from the travel stories on the web and converted into interactive trip plans. Conceptually, it’s people liking what other people created earlier.
According to the Knowledge Graph of the world, the algorithm could search for the best travel experience in the solution space. AI could discover new opportunities, new routes, and new itineraries that were never taken or imagined by humans. The algorithm is capable of measuring the experience, assessing the probability of how close it is to the absolute extremum in the context. In other words, AI is aware of how smart it is — we are building Artificial Smartness. Trip plans made by AI will be shown in the second carousel and could be liked by travelers.
Release Schedule of v18
Most probably we will release the major version without social mechanics. There will be a second carousel of trip plans for several select countries.
The social mechanics such as Liking will be added later and released as the minor version of v18.
There is a backlog. Here are some high-level areas from the backlog.
Trip plan highlights.
Current trip plans created on the Curiosio platform lack human-friendly descriptions, highlights, keywords. We are planning to implement trip highlights together with a UX facelift. The look & feel of the open trip plan is going to change significantly. It is going to improve usability and usefulness. It will be easier to compare the plans and understand the plans. Productivity functions will be added to the plans (Save, Print, Compare).
So far we visualized only three countries: the United States, France, and Italy. Having pictures for more countries is important. We are still experimenting with the imagery — regular pictures, panoramas, paintings, drawings, maps, symbols. Videos, Augmented Reality, Augmented Virtuality are considered conceptually, so we must make decisions about what our platform can/will do in the mid-term.
The next visual country will be the United Kingdom. Most probably it will become public with the v18 major release.
“Last mile” aka fine-tuning.
Fine-tuning features are needed to allow users to finalize almost good trip plans. Sometimes only the user knows what the plan must be. At a large scale, the fine-tuning is Machine Learning play. It is the data for ML. What do travelers want to change in the trip plans produce by the algorithm? What do travelers want to change in the other travelers’ plans?
We consider visuals more important than fine-tuning. One reason: Curiosio is a visual product, the majority of travelers are visuals and demand a picture. Another reason: there are possible hacks to make fine-tuning work even in v17.
- Specify extra travel-through points to stay longer in the area.
- Specify extra travel-through places to stay longer in the city or park.
- Set custom perimeter. Shape the perimeter to the extremes by shaping the browser window.
- Export to Google Maps and do fine-tuning there. Use Google Maps for navigation.
Hacks are hacks — they don’t produce any data for ML but they allow getting more from Curiosio right away.
Thank you for reading down here. Help Ukraine to win the war, stay tuned, and always follow your curiosity!