700M images across 130M users in a week:
GENERATIVE AI
-
Making Chain-of-Thought Monitoring Viable for AI Safety
By
–
To make CoT monitoring a viable way to catch safety issues, we’d need a way to make CoT more faithful, evidence for higher faithfulness in more realistic scenarios, and/or other measures to rule out misbehavior when the CoT is unfaithful. Read the paper: https://
assets.anthropic.com/m/71876fabef0f
0ed4/original/reasoning_models_paper.pdf
… -
Reasoning Models Don’t Reveal When Using Hints
By
–
We slipped problem-solving hints to Claude 3.7 Sonnet and DeepSeek R1, then tested whether their Chains-of-Thought would mention using the hint (if the models actually used it). Read the blog: https://
anthropic.com/research/reaso
ning-models-dont-say-think
… -
Anthropic Research: Reasoning Models Fail Verbalize Accurately
By
–
New Anthropic research: Do reasoning models accurately verbalize their reasoning? Our new paper shows they don't. This casts doubt on whether monitoring chains-of-thought (CoT) will be enough to reliably catch safety issues.
-
Hundreds of Styles Available with Custom Prompt Restyling
By
–
you can choose among hundreds of styles and type a prompt that describes the way how you want the re-style to happen. pic.twitter.com/l0zPmGuv9u
— KREA AI (@krea_ai) 3 avril 2025you can choose among hundreds of styles and type a prompt that describes the way how you want the re-style to happen.
-
Krea AI Launches New Video Generation Tool on Carousel
By
–
you will find this new tool on our new carousel in https://t.co/V31Uw5atb6
— KREA AI (@krea_ai) 3 avril 2025
start by uploading any reference video you want. pic.twitter.com/iSqoihrf59you will find this new tool on our new carousel in http://
krea.ai start by uploading any reference video you want. -
AI Video Re-style Controls: Depth, Edges, Pose Motion Capture
By
–
we offer 4 main controls for Video Re-style.
— KREA AI (@krea_ai) 3 avril 2025
with "Depth", the AI will keep the motion that captures from a depth map on the original video.
"Edges" will capture the motion from the edges on the original video – great for close-ups or faces talking.
"Pose" will only capture… pic.twitter.com/rN0oH5yzn0we offer 4 main controls for Re-style. with "Depth", the AI will keep the motion that captures from a depth map on the original video. "Edges" will capture the motion from the edges on the original video – great for close-ups or faces talking. "Pose" will only capture
-
Video Re-style: AI Tool for Consistent Style Transfer
By
–
introducing Video Re-style.
— KREA AI (@krea_ai) 3 avril 2025
this new tool lets you change the style of any video while keeping the original movement consistent.
learn how 👇 pic.twitter.com/GmfxazMxoNintroducing Re-style. this new tool lets you change the style of any video while keeping the original movement consistent. learn how
-
Convergence AI upgrades Deep Work with parallelisation
By
–
BREAKING 🚨: @convergence_ai_ released an upgrade to their Deep Work feature, which now supports parallelisation of agentic work. pic.twitter.com/RyKw98pVwg
— 🚨 AI News | TestingCatalog (@testingcatalog) 3 avril 2025BREAKING : @convergence_ai_ released an upgrade to their Deep Work feature, which now supports parallelisation of agentic work.
-
AI-Powered Video Search and Cloud Production Tools at NAB Show 2025
By
–
AI-powered video search, real-time captioning across languages, cloud-based production tools, and next-gen encoding are just some of the innovations we're showcasing with partners at @NABShow
. Get a preview of what to expect at #NABShow 2025 in this blog: https://
nvda.ws/3XIN2dJ