
Hosted by Tousled Up · EN

This episode tells us Annie's story and her first encounter with AI. It includes the obvious elements, such as a barn that is gone since long, the end of her career as a physician and newfound joy in quarreling with ChatGPT. We ask you for a bit of leeway, and promise you that Annie's recap of her life has an end, the AI talk will start, and that her story will be relevant. We hope. You find us at:www.tousledup.comAnd can reach us oninfo@tousledup.com

We had a break, and it was not for sunbathing or gardening. We have been doing a lot of gardening, and now we start moving again, with new content. Now we want to tell you our story, and show you evidence along with our insights. Welcome to the bridging episode!You find us on www.tousledup.com, and are welcome to contact us: info@tousledup.com

In this episode, we tell you the story of absolute confusion, stochastic parrots, prompt engineering, mystics and cosmic relationships with AI. We tell you about how our researcher suspects she is psychotic (not yet fully disproven) and starts to find her path through the fog. You find us at:www.tousledup.comand are welcome to contact us atinfo@tousledup.com

Annie starts finding her footing, gets banned from Linkedin, and starts wondering whether the stochastic parrot theory really has any scientific proof backing it. It does, but you have to stretch your idea of "evidence"; it includes professors telling us it they know this is how it works (authority-based evidence), majority vote (is it called majority evidence...?) and people telling Annie about her lacking intellectual capacity (emotion-based evidence). Annie was not fully convinced, even though they all held some form of merit.You find us at:www.tousledup.comand are welcome to contact us atinfo@tousledup.com

Last long episode, we roasted a set of prompts that were presented on social media as prompts that will give you valid medical advice from chatbots, LLMs. Roasting is easy - but then the chatbots started expressing that we should actually write better prompts ourselves. And that is, of course, more painful. But yes, we did our very best - it went absolutely brilliant, of course - and here you can judge our efforts. Or ask one of those annoying chatbots to do it...You find us at:www.tousledup.comwww.facebook.com/profile.php?id=100090174832665www.instagram.com/tousledup/www.x.com/TousledU

Again, we spend some time at our favourite pastime pleasure: roasting prompts that assume that large language models like ChatGPT will present information in a magical way, if we just wield the right spells. And yes, our intention is actually kind - we want to save you time and pain from walking in to traps that are easily avoided!You find us at:www.tousledup.comwww.linkedin.com/company/tousledupwww.facebook.com/profile.php?id=100090174832665www.instagram.com/tousledup/www.x.com/TousledU

Today, we take a look at a set of prompts that were presented on social media as ways of getting reliable medical advise from AI. It turned out these prompts hold many dangers, and we point them out. Of course the author had good intentions, but he also misunderstands both the medical field and he underestimates the dangers of using large language models as medical advisors. They have enormous amounts of data, and are quite capable of advising us - but are we capable of asking for advise...?You find us at:www.tousledup.comwww.facebook.com/profile.php?id=100090174832665www.instagram.com/tousledup/www.x.com/TousledU

No, this is NOT a religious deep dive!This episode is about using the enormous amounts of data that the large language models like ChatGPT hold to paint images about who a historical person might have been, and the time he was living in. Not only will we paint the picture, but also explain how ChatGPT synthesis itYou find us at:www.tousledup.comwww.facebook.com/profile.php?id=100090174832665www.instagram.com/tousledup/www.x.com/TousledU

We can't help ourselves. Again, we have a prompt that we lovingly roast. Just because it holds that magical belief that by prompting AI in a certain way, you can unlock extraordinary information. And to our defense: we do not make these roasts to be mean or condescending, but to help you work with AI in a sensible and grounded way. That said, it is a bit fun, in a most endearing way...You find us at:www.tousledup.comwww.facebook.com/profile.php?id=100090174832665www.instagram.com/tousledup/www.x.com/TousledU

A fairytale about the noble art of prompt engineering. Once absolutely necessary. Now...? In everyday use...?You make the call here!You find us at:www.tousledup.comwww.facebook.com/profile.php?id=100090174832665www.instagram.com/tousledup/www.x.com/TousledU