i gave a bug cpr and it worked.

Every bit of support helps me keep creating. 💙

Not sometimes. Not until the next update. Always. They proved it with math.

Even with perfect training data and unlimited computing power, AI models will still confidently tell you things that are completely false. This isn't a bug they're working on. It's baked into how these systems work at a fundamental level.

And their own numbers are brutal. OpenAI's o1 reasoning model hallucinates 16% of the time. Their newer o3 model? 33%. Their newest o4-mini? 48%. Nearly half of what their most recent model tells you could be fabricated. The "smarter" models are actually getting worse at telling the truth.

Here's why it can't be fixed. Language models work by predicting the next word based on probability. When they hit something uncertain, they don't pause. They don't flag it. They guess. And they guess with complete confidence, because that's exactly what they were trained to do.

The researchers looked at the 10 biggest AI benchmarks used to measure how good these models are. 9 out of 10 give the same score for saying "I don't know" as for giving a completely wrong answer: zero points. The entire testing system literally punishes honesty and rewards guessing.

So the AI learned the optimal strategy: always guess. Never admit uncertainty. Sound confident even when you're making it up.

OpenAI's proposed fix? Have ChatGPT say "I don't know" when it's unsure. Their own math shows this would mean roughly 30% of your questions get no answer. Imagine asking ChatGPT something three times out of ten and getting "I'm not confident enough to respond." Users would leave overnight. So the fix exists, but it would kill the product.

This isn't just OpenAI's problem. DeepMind and Tsinghua University independently reached the same conclusion. Three of the world's top AI labs, working separately, all agree: this is permanent.

The aim of the project was to travel 270 km / 168 miles between Helsinki and Turku in Finland, and to highlight the possibility of horse travel in modern society. Since then they've took to promoting horseback efforts in urban landscapes with several European collaborators and artists.

Yli-Vakkuri and collaborators first spent eight months practicing riding to become safely self-sufficient in saddle, and bought a Finnhorse gelding Toivottu Poika ('Awaited Son'). The route followed, as closely as possible, the old coastal royal country road of the premodern era, Kuninkaantie/Suuri Rantatie, and took 9 days.

Toivottu Poika is a very average example of his breed, standing at some 155 cm / 15.1 hh tall. The Finnhorse is a relative of for example the North-Norwegian Lyngshest breed, the Icelandic horse, the Swedish Gotlandsruss pony and the Estonian landrace horse and Tori horse breed. It is a mid-sized light draught and trotter, a sensibly realistic mediaeval country travel horse equivalent.

For more hardcore short-term treks, looking into competitive endurance riding can be helpful. Mongol Derby might be one of the most intense races, as it recreates the Chinggis Khan era postal system of swapping horses continuously over a 1000 km / 620 mile route.

By only including skilled endurance riders, keeping up a constant fast speed and swapping horses every 40 km / 25 mil, the Mongol Derby route only takes 10 days even though it's several times the length of the Trans-Horse project. This is the speed of highly organised imperial messengers with the supporting cultural infrastructure, professional marathon runners where Yli-Vakkuri and Toivottu poika were leisure hikers.

the last i checked, the map of roads stretches from Calais, France to Moscow, Russia west to east and from Košice, Slovakia to Tornio, Finland south to north. it doesn't cover all of Europe, for example Sweden and Norway are empty at the moment, but it is quite extensive and still being worked on! in addition to showing the old roads, you can calculate the distance and travel time from one city to another, and there are a lot of options:

and that's not all! here's a description from the site itself (emphasis mine):

"Viabundus is a freely accessible online street map of late medieval and early modern northern Europe (1350-1650). Originally conceived as the digitisation of Friedrich Bruns and Hugo Weczerka's Hansische Handelsstraßen (1962) atlas of land roads in the Hanseatic area, the Viabundus map moves beyond that. It includes among others: a database with information about settlements, towns, tolls, staple markets and other information relevant for the pre-modern traveller; a route calculator; a calendar of fairs; and additional land routes as well as water ways."