rewon
101 posts

rewon
@rewonfc
Here for the research papers
San Francisco, CA Katılım Kasım 2012
290 Takip Edilen1.1K Takipçiler

Yesterday, my wife and I laid to rest our cat son, Darth. Our hearts are hurting. We will always miss and love him.
Two years ago, he was diagnosed with osteosarcoma in his front leg. We had his limb removed, and he thrived as a hopping tripod. A year later, in his annual chest x-ray, we were shocked to learn his cancer had metastasized and that he had 4-6 weeks. We vowed to make every day he had left as peaceful and loving as possible.
Out of sheer force of will and love, Darth fought his very aggressive cancer without medical intervention to give his family a whole extra year with him. His illness won't define him. His mom and I will always remember him for his incredible sweetness, his hilarious hiding abilities while reappearing only for his favorite humans, and his love for his yellow catnip bananas, grandma's lettuce, high places, cheek massages, and for always being nearby.
We are grateful to be his parents.
English

Personal news: I'm joining @AnthropicAI! 😄 Anthropic's approach to AI development resonates significantly with my own beliefs; looking forward to contributing to Anthropic's mission of developing powerful AI systems responsibly. Can't wait to work with their talented team, including a number of great ex-colleagues from OpenAI and Google, and tackle the challenges ahead!
English

Oh! Officially a published first author at a major AI conference. It’s my first “real real paper” (miniF2F was @KunhaoZ and GPTf is unpublished), so I’m kind of psyched.
openreview.net/pdf?id=-P7G-8d…
🤙@jessemhan @KunhaoZ @ibab_ml @ilyasut
Best co-authors.
English
rewon retweetledi

Excited to announce that we’ve raised $1.3B to build one of the largest clusters in the world and turbocharge the creation of Pi, your personal AI.
forbes.com/sites/alexkonr…
English
rewon retweetledi

It’s a big week! We’ve raised $1.3 billion and are building the world’s largest AI cluster (22k H100s).
We’re grateful for our investors and new funding that will help us accelerate our mission to make personal AI available to every person in the world. inflection.ai/inflection-ai-…
English
rewon retweetledi

We have amazing results to announce! Inflection-1 is our new best-in-class LLM powering Pi, outperforming GPT-3.5, Llama and PALM-540B on major benchmarks commonly used for comparing LLMs. inflection.ai/inflection-1
English

I gave my dissertation talk last month! 🥳 Thanks to my advisor Tom (@cocosci_lab), my committee, and the communities at Berkeley, Princeton, and beyond for fun and support along the way ❤️

English

“Golden retriever riding in a hot air balloon”
DALLE 2 didn’t get the semantics fully right, but I might like this idea more
#dalle2

English

A sign of how bad Google search has become: the page containing the original quote is not even on the first page of search results for it.
google.com/search?q=%22Pr…
Bing does better though:
bing.com/search?q=%22Pr…
English

I defended today! Thanks very much to my advisors @boazbaraktcs & Madhu Sudan, and all of my friends & collaborators along the way.
In the Fall I'll be joining UCSD as a postdoc with Misha Belkin. If you're interested in my research and want to collab, do reach out!

English
rewon retweetledi
rewon retweetledi

I've ported @rewonfc's very deep VAE (arxiv.org/abs/2011.10650) from PyTorch to JAX/Flax! Hope other JAX users find this SOTA VAE useful as a forkable baseline... github.com/j-towns/vdvae-….
English
rewon retweetledi

My post on SotA image generative models was released 🥳
Featured 7 notable recent papers with emphasis on:
- VD-VAE
- VAE + discriminator (e.g. VQGAN, DC-VAE)
- Diffusion models (e.g. DDPMv2)
Plus some notes on scaling (e.g. DALL-E) and evaluation.
arankomatsuzaki.wordpress.com/2021/03/04/sta…
English
rewon retweetledi

IMO, best empirical proof to date that AI can be creative. After this sinks in, will there be any naysayers left?
Durk Kingma@dpkingma
"The images are preprocessed to 256x256 resolution during training. [...] each image is compressed to a 32x32 grid of discrete latent codes using a discrete VAE that we pre-trained using a continuous relaxation." GPT + VAE + scale = impressive results! openai.com/blog/dall-e/
English

Congrats to Aditya and the rest of the team for an awesome release!
Ilya Sutskever@ilyasut
Synthetic capybaras in different styles openai.com/blog/dall-e/
English




