Mike Tamir, PhD
(99+) Introducing Bloom: The Generative Conversion Network for Product Photos | LinkedIn https://t.co/yRFx8Ew8Aq
tweet

elvis
Mathematical Capabilities of ChatGPT

Was just thinking about a similar idea after the ChatGPT update yesterday.

Nice to see other researchers are also thinking about investigating this more in-depth.

https://t.co/m0N72cC8lN https://t.co/KidvA6KUTv
tweet

AK
Scaling laws for single-agent reinforcement learning

abs: https://t.co/6SgKgClVab https://t.co/DUfQLTWnBr
tweet

Earth Is A Sales Funnel For SATAN
The world is a forest which from time to time must be set on fire for the good of all life
tweet

AK
The Flan Collection: Designing Data and Methods for Effective Instruction Tuning

abs: https://t.co/zwh87UyH0N
github: https://t.co/vQNwTn628K https://t.co/jHJld4sb5q
tweet

surea.i
thinking about what my profile pic might look like in an alternate timeline. made with @layer_ai

(only time I've ever changed it was that time I started a twitter gang-impersonation of @takyon236 )
#aiart #aiia #imagesynthesi #synthography https://t.co/E8lO8iJw0m
tweet

AK
Grounding Language Models to Images for Multimodal Generation

abs: https://t.co/kywcCjNTHm
project page: https://t.co/aKz2FpMYzQ

method to ground pretrained text-only LMs to the visual domain, enabling them to process and generate arbitrarily interleaved image-and-text data https://t.co/m97W8XNza8
tweet

Aran Komatsuzaki
Faithful Chain-of-Thought Reasoning

Faithful CoT outperforms CoT on 9 out of 10 reasoning datasets by decomposing a reasoning task into two stages.

https://t.co/B1xiiVFoMZ https://t.co/5tyosFRK8i
tweet

UnknownCollector - Auction live ❤️
RT @agugn_: "Have no fear of perfection--you'll never reach it."
--Salvador Dali

Embrace Fear, enjoy the process

Available to mint until 3 Feb
“Joyful Fear”, OE on @manifoldxyz
ERC1155

GM & have a nice day

Link below ⚡️💀🙏🏽🌷 https://t.co/CM0YLuHLti
tweet

Dave Craige
*working to keep all 16 members of the @TheStudio_io core team updated and aligned.

Twice a month team updates in our Discord. https://t.co/IYkt3t4R8n
tweet

AK
FLAME: A small language model for spreadsheet formulas

abs: https://t.co/tqTBzET7xW

FLAME(60M) can outperform much larger models, such as Codex-Davinci (175B), Codex-Cushman (12B), and CodeT5 (220M), in 6 out of 10 settings https://t.co/pGIFNM5d0g
tweet