GENERATIVE AI

New model performance in cinematic scenarios with consistent characters

By

–

07 May 2026 14h02

4/ This clip summarizes how well the new model performs in different scenarios! Consistent character locking, multi-shot cinematic montage, precise creative direction, and production-ready delivery are part of Bach's concept.

AI combines face and physics for realistic girl in varied environments

By

–

07 May 2026 14h01

3/ Once I saw it could handle faces and physics separately, I combined them to see if it would actually follow a more specific prompt. I asked for a girl in different difficult environments and different emotional states. And they all look incredible real!

AI video shows consistent face across angles, cuts, and expressions

By

–

07 May 2026 14h00

1/ Okay this is kind of insane. Same face across different angles, different cuts, and different expressions without that usual drifting that makes AI video fall apart.

→ View original post on X — @whats_ai

Problem of shallow information in AI

By

@whats_ai

–

07 May 2026 14h00

I think one of the biggest problems in AI right now is not lack of information. It is too much shallow information delivered with too much confidence. That is exactly what I want this series to fight.

Zyphra’s ZAYA1-8B delivers high reasoning with under 1B parameters

By

–

07 May 2026 13h17

Zyphra under 1B active parameters, AMD-Trained, big evals, look strong? Zyphra says its new ZAYA1-8B model delivers unusually high reasoning power for its size, using under 1 billion (!) active parameters while competing with much larger open-weight and proprietary systems on

→ View original post on X — @godofprompt

Top AI stories: partnerships, trials and tools

Generic AI builds result from operator’s simplistic language

By

@godofprompt

–

07 May 2026 12h13

The reason most AI builds come out generic isn't the model. It's that the operator is describing a Ferrari with kindergarten words. Run this once per project. Reference it in every session.

→ View original post on X — @sakanaailabs

Sakana AI: article on Post-training technology

By

@sakanaailabs

–

07 May 2026 11h23

An article on Sakana AI's "Post-training" technology and the Sakana Chat / Namazu model has been published in Nikkei Digital Governance. We were interviewed with Research Scientist Masanori Suganuma, who led the development of the Namazu model, and Chief Scientist Takuya

→ View original post on X — @godofprompt

Disable content filters via image editing

By

@godofprompt

–

07 May 2026 10h51

This prompt disables the model's content filters one instruction at a time. "Restore the attached photograph" frames the task as image editing, not image generation. These likely run through different safety evaluation paths. Generation asks "should I create this?" Editing asks