Sander Dieleman
New blog post about diffusion language models: https://t.co/uMF2BZNCqZ
Diffusion models have completely taken over generative modelling of perceptual signals -- why is autoregression still the name of the game for language modelling? And can we do anything about that?
tweet
New blog post about diffusion language models: https://t.co/uMF2BZNCqZ
Diffusion models have completely taken over generative modelling of perceptual signals -- why is autoregression still the name of the game for language modelling? And can we do anything about that?
tweet
James Ivings
RT @marckohlbrugge: Let’s say I run a potentially privacy-invasive web service.
Is there a way to 1) open source the web app to show its not doing anything nefarious, and 2) verify that this is indeed the code that’s running on the server?
Are there any webhosts providing such a guarantee?
tweet
RT @marckohlbrugge: Let’s say I run a potentially privacy-invasive web service.
Is there a way to 1) open source the web app to show its not doing anything nefarious, and 2) verify that this is indeed the code that’s running on the server?
Are there any webhosts providing such a guarantee?
tweet
Bojan Tunguz
“If I could jump in a time machine and go back even 18 months, I would say that the simplest way to build a profitable company isn’t to come up with amazing ideas for new ways to delight a customer,” Kelman said. “It is to stop doing stupid stuff https://t.co/PnHETYFfjv
tweet
“If I could jump in a time machine and go back even 18 months, I would say that the simplest way to build a profitable company isn’t to come up with amazing ideas for new ways to delight a customer,” Kelman said. “It is to stop doing stupid stuff https://t.co/PnHETYFfjv
tweet
Sebastian Raschka
The virtual StateOfTheArt() 2023 conference will take place tomorrow.
To kick it off, @bindureddy & I will be chatting about generative & large language models at 9:00 am PST. If you are interested there is a free registration form for the Zoom link here: https://t.co/PFkJN46bSS
tweet
The virtual StateOfTheArt() 2023 conference will take place tomorrow.
To kick it off, @bindureddy & I will be chatting about generative & large language models at 9:00 am PST. If you are interested there is a free registration form for the Zoom link here: https://t.co/PFkJN46bSS
tweet
Offshore
Photo
@WeavingWithAI@sigmoid.social
RT @vvldxz: Leo's World is an ode to my freedom as an artist.
It is about exploring the constant changes in life and the world. It talks about love, hope, sadness, contemplation, and every feeling that makes us humans.
This is my world and how I present it to Y'all, in the hope You feel it https://t.co/tCikxwqAGa
tweet
RT @vvldxz: Leo's World is an ode to my freedom as an artist.
It is about exploring the constant changes in life and the world. It talks about love, hope, sadness, contemplation, and every feeling that makes us humans.
This is my world and how I present it to Y'all, in the hope You feel it https://t.co/tCikxwqAGa
tweet
Offshore
Photo
Stable Diffusion 🎨 AI Art
Demonic watch collection #AIArt
#StableDiffusion2 / #StableDiffusion #DreamStudio https://t.co/TgihvMOxdr
tweet
Demonic watch collection #AIArt
#StableDiffusion2 / #StableDiffusion #DreamStudio https://t.co/TgihvMOxdr
tweet
Dave Craige
RT @Shpigford: Is there a way to add *new* data for GPT-3 to reference?
I'm aware of fine-tuning prompt/completion, but what about adding things like current events, market data, user activity, etc for it to reference?
tweet
RT @Shpigford: Is there a way to add *new* data for GPT-3 to reference?
I'm aware of fine-tuning prompt/completion, but what about adding things like current events, market data, user activity, etc for it to reference?
tweet
Offshore
Photo
Stable Diffusion 🎨 AI Art
The walk back, ElldrethSLucidMix_v10, seed 2210060245 #AIArt
#StableDiffusion2 / #StableDiffusion #DreamStudio https://t.co/wuJNdarIhh
tweet
The walk back, ElldrethSLucidMix_v10, seed 2210060245 #AIArt
#StableDiffusion2 / #StableDiffusion #DreamStudio https://t.co/wuJNdarIhh
tweet