LogoThread Easy
  • 発見
  • スレッド作成
LogoThread Easy

Twitter スレッドの万能パートナー

© 2025 Thread Easy All Rights Reserved.

探索

Newest first — browse tweet threads

Keep on to blur preview images; turn off to show them clearly

RT @littmath: OK, I think GPT 5.2 Pro is actually a step change in usefulness for my applications (algebraic geometry/number theory researc…

RT @littmath: OK, I think GPT 5.2 Pro is actually a step change in usefulness for my applications (algebraic geometry/number theory researc…

I work on AI at OpenAI. Former VP AI and Distinguished Scientist at Microsoft.

avatar for Sebastien Bubeck
Sebastien Bubeck
Tue Dec 16 16:59:07
Just trying out DoorDash’s new AI food recs app! It’s pretty well-built — congrats to them!

Surprisingly, it lacks the funnel for food delivery services like DoorDash

I wonder what’s their end goal for this… I know they’ve been trying to branch out to IRL dining lately

Just trying out DoorDash’s new AI food recs app! It’s pretty well-built — congrats to them! Surprisingly, it lacks the funnel for food delivery services like DoorDash I wonder what’s their end goal for this… I know they’ve been trying to branch out to IRL dining lately

“The woman scooping Silicon Valley” — BBC・hacker turned builder, blogger & consultant・ex: Threads, Instagram, etc

avatar for Jane Manchun Wong
Jane Manchun Wong
Tue Dec 16 16:59:04
incredible and not surprised.
still human-in-the-loop but gpt 5+ series has repeatedly shown itself to be an amazing research partner.

i strongly recommend checking out the thread. 
clearly current path of scaling has yielded “nothing beyond simple automation”.

incredible and not surprised. still human-in-the-loop but gpt 5+ series has repeatedly shown itself to be an amazing research partner. i strongly recommend checking out the thread. clearly current path of scaling has yielded “nothing beyond simple automation”.

making models learn • eXperiments lab

avatar for tokenbender
tokenbender
Tue Dec 16 16:58:34
@heyglif You need images that show each stage of the "renovation." 

This means using an image edit model to change one thing at a time.

I'd recommend picking a few of the following: floors, ceiling, paint, furniture, rug, wall art, lighting. 

These were my steps for the above video ⬇️

@heyglif You need images that show each stage of the "renovation." This means using an image edit model to change one thing at a time. I'd recommend picking a few of the following: floors, ceiling, paint, furniture, rug, wall art, lighting. These were my steps for the above video ⬇️

The annoying part of this workflow is prompting the video model so many times with each start and end frame - and then stitching the videos together. Luckily, the new @heyglif agent can do this for you. AND it can make the progressive image frames. Use this prompt: "I want to make a fun renovation video. Take my starting image and generate each stage of the renovation (flooring, walls, ceiling, rug, decor). Use Kling to animate creative transitions. And then stitch the videos together." https://t.co/3lRSx5TNlJ

avatar for Justine Moore
Justine Moore
Tue Dec 16 16:57:12
RT @PKirgis: In our most recent evaluations at @halevals, we found Claude Opus 4.5 solves CORE-Bench. How? 

Opus 4.5 solves CORE-Bench bec…

RT @PKirgis: In our most recent evaluations at @halevals, we found Claude Opus 4.5 solves CORE-Bench. How? Opus 4.5 solves CORE-Bench bec…

Princeton CS prof and Director @PrincetonCITP. Coauthor of "AI Snake Oil" and "AI as Normal Technology". https://t.co/ZwebetjZ4n Views mine.

avatar for Arvind Narayanan
Arvind Narayanan
Tue Dec 16 16:56:28
Here is a rather pretty configuration using this feature, from the community.

Here is a rather pretty configuration using this feature, from the community.

Working on a new terminal: Ghostty. 👻 Prev: founded @HashiCorp. Created Vagrant, Terraform, Vault, and others. Vision Jet Pilot. 👨‍✈️

avatar for Mitchell Hashimoto
Mitchell Hashimoto
Tue Dec 16 16:55:37
  • Previous
  • 1
  • More pages
  • 609
  • 610
  • 611
  • More pages
  • 5634
  • Next