we are releasing two versions of our foundation image model: Krea 2 Raw and Krea 2 Turbo. combined, they bring unprecedented opportunities for fine-tuning, post-training, and working with diverse aesthetics. the core idea is: train on Raw, generate with Turbo.
AI
-
Krea 2 Raw: Undistilled Model for Further Development
By
–
Krea 2 Raw is an undistilled model from mid-training stage that we created while training Krea 2 Medium. The lack of distillation, fine-tuning, or post-training make it an ideal model to build on.
-
Krea 2 open weights: Raw and Turbo models released
By
–
today, we release the open weights of Krea 2.
— Krea (@krea_ai) 23 juin 2026
welcome Krea 2 Raw and Krea 2 Turbo, an undistilled model from mid-training meant to be fine-tuned, and a fast distilled version with a wide aesthetic diversity.
read the details below 👇 pic.twitter.com/3ymzUL2bxvtoday, we release the open weights of Krea 2. welcome Krea 2 Raw and Krea 2 Turbo, an undistilled model from mid-training meant to be fine-tuned, and a fast distilled version with a wide aesthetic diversity. read the details below
-
Centralized AI processing limits responsiveness for time-sensitive use cases
By
–
Most AI systems still rely on centralized processing: → Data is sent → Processed → Returned That model works, but for time-sensitive use cases, responsiveness and control become critical.
-
AI Limitations: Latency and Decision Location Drive Edge
By
–
AI isn’t failing because it lacks intelligence.
— Ronald van Loon (@Ronald_vanLoon) 23 juin 2026
In many cases, it’s limited by latency and where decisions happen.
For time-sensitive environments, that’s a critical constraint.
Here’s why AI is moving to the edge, and what it unlocks…
@TMobileBusiness Partner pic.twitter.com/XmoiFQQscvAI isn’t failing because it lacks intelligence. In many cases, it’s limited by latency and where decisions happen.
For time-sensitive environments, that’s a critical constraint. Here’s why AI is moving to the edge, and what it unlocks… @TMobileBusiness Partner -
NVIDIA brings reliable AI agents 24/7 to telecom operations
By
–
At #DTWIgnite, in partnership with our partners, we present the data, models, simulation, and secure execution stack enabling telecom operators to build flows of
-
Serious autonomous runtime generates departments of agents, not just a joke
By
–
the @flowith video is pretty hilarious, putting toilet paper on empty desks to joke about zero employees.
— Charly Wargnier (@DataChaz) 23 juin 2026
But don't be fooled.
I've seen demos and their tech is serious:
> an autonomous runtime generating entire departments of agents for multi-step tasks that really works 🦾 pic.twitter.com/Zx42TJEbnXthe @flowith video is pretty hilarious, putting toilet paper on empty desks to joke about zero employees. But don't be fooled. I've seen demos and their tech is serious: > an autonomous runtime generating entire departments of agents for multi-step tasks that really works
-
Disappointment: Sonnet 5 soon, GPT-5.6 postponed
By
–
What a disappointment. So only Sonnet 5 soon. GPT-5.6 postponed.
-
Testing AI bug detection by editing out bugs for false positive check
By
–
Testing this more. I think it's doing the right thing based on the image rather than prior knowledge. In these two tests I edited the original image to remove the bugs while keeping everything else the same. I was looking for a false positive if it wasn't doing the correct
-
OPENCLAW Tutorial: Installation, Configuration, Skills and AI Usage
By
–
NEW VIDEO in the LAB! Now that during the holidays you will have more free time, I bring you a tutorial on one of the most important AI macro-trends of the year: OPENCLAW Installation, configuration, Skills, AI usage in