Dreaming on dreams or designs?

Do they pass the "reality threshold"? How do we know?

We’re back with another edition of Machine Learnings, brought to you by the folks at Heyday.

Heyday is an AI-powered memory assistant that resurfaces content you forgot about while you browse the web.

This week we’re sharing part groundbreaking theory, part practical uses for AI. At the end of the day, AI is still software, and people don’t use software that doesn’t provide value. This is one reason the hype cycle seems to have tuned down a touch.

We’re building with AI and writing lots about coaching these days. If you work with – or are – an executive coach, Heyday’s got a special experience for you. Let’s talk.

-@samdebrule

---

What we're reading.

1/ The challenge of benchmarking video generation tools is multifaceted, and a team of Google researchers is taking it head-on. Meet StoryBench. Learn more at Papers with Code >

2/ Douglas Hofstadter’s Gödel, Escher, Bach won him a general nonfiction Pulitzer Prize in 1980. Today he’s breaking down where AI gets it right and wrong, and where he’s hoping the tech takes us. Learn more at The Atlantic >

3/ A new paper has signaled the start of reinforced self training LLMs. What does that mean for us makers? Learn more at arXiv >

4/ How do we determine what’s real or fake, down to the neurological function in our brain? Enter the “reality threshold”. Learn more at WIRED >

5/ Google Labs are putting out some amazing experiments. This one is built for rappers, writers and wordsmiths, with the help of the great Lupe Fiasco. Learn more at With Google >

6/ Dreaming. It’s useful beyond our imaginations, and it’s why I have an inkling of hope for this new Snapchat product even if the first version looks elementary. Learn more at TechCrunch >

7/ The highest performing LLM known to mankind today. Learn more at Phind >

---

Research for this edition of Machine Learnings was enhanced by Heyday, the AI-powered memory assistant.