Why would I use a low fidelity version when I can just recall a high fidelity version directly in my mind? You too can achieve this by training dream recall, by keeping a dream journal for example. After a while you will notice you can remember nearly everything about dreams and will start lucid dreaming as well.

Waking up every morning to relay my dream of Will Smith eating spaghetti once again.

For real though, who exactly is this for? People who want to see an AI's take on the real dream they just had?

I think it's meant as an art piece, not a product. I liked that the call to action is "BUILD YOUR OWN" instead of "BUY NOW ($599)".

This makes no sense. If I'm putting my dreams into words, that's already as good as it's gonna get. Why make it even lossier by involving AI? In fact, it's even worse than just lossy since incorrect details will be added in!

This idea is also pretty obvious. Who hasn't tried describing a dream to AI and been disappointed with the slop it generates? It will never look anything like what was imagined. Most of the imagery in my mind is unique to stuff I have experienced in my real life offline. AI training data is biased very far away from anything so candid, and if the images are wrong then they also cannot convey the same emotions that the words did.

This problem occurs again and again with damn near everything AI generates. All emotion and style are replaced with the kind of stale and cold feelings you only get from stock photos, trashy low effort music, and corporate speak.

It's a fancy dream journal. It doesn't record brain activity.

We will one day have the technology to actually record dreams, however. See https://pubmed.ncbi.nlm.nih.gov/40876481/, Reconstructing high-resolution visual perceptual images from human intracranial electrocorticography signals (2025)

> Reconstruction of visual perception from brain signals has emerged as a promising research topic. Electrocorticography (ECoG) is a kind of high-quality intracranial signal with good spatiotemporal resolution that offers some new opportunities. However, according to our knowledge, there are no studies to reconstruct the perceived images from human ECoG signals at present.

> We have conducted the pioneering work and developed a novel pipeline that integrates Talairach coordinate alignment masked autoencoders (TA-MAE) with denoising diffusion probabilistic models. Our approach exploits the spatiotemporal dynamics of human ECoG signals, enabling the restoration of details in high-resolution