Aella
last year i tried to estimate "what % of sex workers in the US are trafficked?"
Here's a breakdown of how I arrived at that estimate:
https://t.co/4AGxWxS9Jc
tweet
last year i tried to estimate "what % of sex workers in the US are trafficked?"
Here's a breakdown of how I arrived at that estimate:
https://t.co/4AGxWxS9Jc
tweet
Offshore
GIF
pugson
RT @0xca0a: new in drei <MeshTransmissionMaterial> for more convincing glass/gelatin/epoxy transmission w/ RGB shift & noise. based on a material that was open sourced by @ore_ukonpower and @junni_jp, w/ some extras added.
<mesh>
<MeshTransmissionMaterial />
demo: https://t.co/FcyJANBCBT https://t.co/95vKSOIdLc
tweet
RT @0xca0a: new in drei <MeshTransmissionMaterial> for more convincing glass/gelatin/epoxy transmission w/ RGB shift & noise. based on a material that was open sourced by @ore_ukonpower and @junni_jp, w/ some extras added.
<mesh>
<MeshTransmissionMaterial />
demo: https://t.co/FcyJANBCBT https://t.co/95vKSOIdLc
tweet
Offshore
Photo
Jim Fan
Happy 7th birthday @OpenAI! Most people know about this company through DALLE & ChatGPT.
I interned at OpenAI in 2016-17 on the “Universe” initiative, which inspired my Minecraft work @NVIDIAI. In its early days, OpenAI hit 7 cool milestones in games, agents & robotics.
Long🧵: https://t.co/T5KD2V3qTK
tweet
Happy 7th birthday @OpenAI! Most people know about this company through DALLE & ChatGPT.
I interned at OpenAI in 2016-17 on the “Universe” initiative, which inspired my Minecraft work @NVIDIAI. In its early days, OpenAI hit 7 cool milestones in games, agents & robotics.
Long🧵: https://t.co/T5KD2V3qTK
tweet
Offshore
Photo
Bojan Tunguz
Redesigning my toilet using the most advanced GPUs I have around. #RTXOn https://t.co/i3BUBfJrOK
tweet
Redesigning my toilet using the most advanced GPUs I have around. #RTXOn https://t.co/i3BUBfJrOK
tweet
Fei Xia
RT @denny_zhou: Discovering and understanding emergent abilities of LLMs, like CoT reasoning, is the most exciting and challenging topic in LLM research. These emergent abilities are not intentionally built by anyone who trained LLMs via predicting next tokens.
tweet
RT @denny_zhou: Discovering and understanding emergent abilities of LLMs, like CoT reasoning, is the most exciting and challenging topic in LLM research. These emergent abilities are not intentionally built by anyone who trained LLMs via predicting next tokens.
tweet
Tyler Richards
Just finished year 1 at @streamlit, what a fun and meaningful way to spend time. Here's to the next one!
tweet
Just finished year 1 at @streamlit, what a fun and meaningful way to spend time. Here's to the next one!
tweet
Offshore
Photo
Lior⚡
A much needed paper. GPT-family models can be pruned to 50% sparsity in one-shot, without any retraining, at minimal loss of accuracy:
- Achieves 60% sparsity on OPT-175B and BLOOM-176B
- 100 billion weights can be ignored at inference time
📄 Paper: https://t.co/GqKKfEQ9SV https://t.co/wbX4PiCKOo
tweet
A much needed paper. GPT-family models can be pruned to 50% sparsity in one-shot, without any retraining, at minimal loss of accuracy:
- Achieves 60% sparsity on OPT-175B and BLOOM-176B
- 100 billion weights can be ignored at inference time
📄 Paper: https://t.co/GqKKfEQ9SV https://t.co/wbX4PiCKOo
tweet
Bojan Tunguz
Orin is about to hit the road. Bigly.
@Nvidia, Foxconn Partner on Electric Vehicles - The Wall Street Journal https://t.co/Q7X8stzeav
tweet
Orin is about to hit the road. Bigly.
@Nvidia, Foxconn Partner on Electric Vehicles - The Wall Street Journal https://t.co/Q7X8stzeav
tweet