A clever way to show how well the upcoming DeepFloyd IF model handles text generation.
#GenerativeAI about to make big progress. https://t.co/qmCODUSvyi

For people that loved the potential of #riffusion (using spectrograms with #stablediffusion): Google Research has published MusicLM, an #AI model that can generate high quality music from a text prompt. You have to listen to the examples on this page:
https://t.co/jzp9oDMapm https://t.co/KCRV4WIbfW

@_Lost_letters_ Some Leonora Carrington in this one.

RT @hardmaru: @StableDiffusion Brutalist Cat

#StableDiffusion2 #AIart https://t.co/n4aY3Cnq1W

Writing:
– time dedicated to thinking: 99%
– time dedicated to fact checking: 1%

Writing with #chatGTP:
– time dedicated to thinking: 1%
– time dedicated to fact checking: 99%

20+ years of career in enterprise IT teach you a lot of things. One of them is that sales teams in big tech vendors are amazing. They do the hardest job, they are the only ones that truly understand the client (because they have the relationship), and yet they are rarely heard.

You often hear the pejorative expression “they are coin operated”, but my respect and admiration for the job that sales team do has only grown my last 13 years I. Gartner and Red Hat.

You, instead, have created 10 slides in PowerPoint full of bullet points nobody will ever read, played Candy Crush, and tweeted that it’s a shame you have to pay $8 for a service that you used for free for over 10 years. https://t.co/jPJKFQDLyA

1/ Tonight I met another successful co-founder. He’s great and I wish him overwhelming success. Every founder believes he/she has an unmatchable competitive advantage against the rivals in the same space. Founders have to believe that to face the incredible challenges of building

2/ a startup. And sometimes it’s true. But the existential threat comes from companies that solve the problem you’re solving from a completely different angle. Not from a company that solves the problem incrementally better.