探索 | Thread Easy - Twitterスレッドを展開 | リーダー・要約・作成

Microsoft just released Fara-7B for CUA (Computer use automation) - Based on Qwen 2.5 VL (Chinese model) - Tested on Ubuntu 24.04.3 LTS (Linux OS) - Trained/Tested on NVIDIA hardware https://t.co/zXsm5EVXwZ

All Things Engineering. Electrical, Mechanical, Software, Firmware, AI, Security and everything in between. Specialize in custom HW/FW/SW for motor control

Engineering Randomness

Tue Nov 25 19:20:01

@bfl_ml I've been running it on @fal - so I was very happy to have a picture to test with @gorkem, @nina_gerov, @robrombach, and @andi_blatt 😅 I'll be sharing results of LoRA training soon. In the meantime - worth checking out this chart! 👀

Partner @a16z AI 🤖 and twin to @omooretweets | Investor in @elevenlabsio, @krea_ai, @bfl_ml, @hedra_labs, @wabi, @WaveFormsAI, @ViggleAI, & more

Justine Moore

Tue Nov 25 19:14:12

Guys you will never believe what I saw when getting an uber from the airport this morning

UI/UX Designer ϟ Prev: lead designer at @super_ ✱ UI inspiration @DamnGoodUI

Josh Millgate

Tue Nov 25 19:10:50

RL is a whole different world. Top 9 LLMs total 2.4T tokens ~= 10 TB. PufferLib trains >1000 TB of observations on a single 6x4090 tinybox. So much work on scaling compute, so little on scaling data!

I build sane open-source RL tools. MIT PhD, creator of Neural MMO and founder of PufferAI. DM for business: non-LLM sim engineering, RL R&D, infra & support.

Joseph Suarez 🐡

Tue Nov 25 19:10:12

“From 2012 to 2020, it was the age of research. From 2020 to 2025, it was the age of scaling. Is the belief that if you just 100x the scale, everything would be transformed? I don't think that's true. It's back to the age of research again, just with big computers.” @ilyasut

Host of @dwarkeshpodcast https://t.co/3SXlu7fy6N https://t.co/4DPAxODFYi https://t.co/hQfIWdM1Un

Dwarkesh Patel

Tue Nov 25 19:10:00

“From 2012 to 2020, it was the age of research. From 2020 to 2025, it was the age of scaling. Is the belief that if you just 100x the scale, everything would be transformed? I don't think that's true. It's back to the age of research again, just with big computers.” @ilyasut

Host of @dwarkeshpodcast https://t.co/3SXlu7fy6N https://t.co/4DPAxODFYi https://t.co/hQfIWdM1Un

Dwarkesh Patel

Tue Nov 25 19:10:00

探索

Newest first — browse tweet threads

探索

Newest first — browse tweet threads

Microsoft just released Fara-7B for CUA (Computer use automation) - Based on Qwen 2.5 VL (Chinese model) - Tested on Ubuntu 24.04.3 LTS (Linux OS) - Trained/Tested on NVIDIA hardware https://t.co/zXsm5EVXwZ

@bfl_ml I've been running it on @fal - so I was very happy to have a picture to test with @gorkem, @nina_gerov, @robrombach, and @andi_blatt 😅 I'll be sharing results of LoRA training soon. In the meantime - worth checking out this chart! 👀

Guys you will never believe what I saw when getting an uber from the airport this morning

RL is a whole different world. Top 9 LLMs total 2.4T tokens ~= 10 TB. PufferLib trains >1000 TB of observations on a single 6x4090 tinybox. So much work on scaling compute, so little on scaling data!

“From 2012 to 2020, it was the age of research. From 2020 to 2025, it was the age of scaling. Is the belief that if you just 100x the scale, everything would be transformed? I don't think that's true. It's back to the age of research again, just with big computers.” @ilyasut

“From 2012 to 2020, it was the age of research. From 2020 to 2025, it was the age of scaling. Is the belief that if you just 100x the scale, everything would be transformed? I don't think that's true. It's back to the age of research again, just with big computers.” @ilyasut