/7 Standard metrics aren't enough to prove an AI remembers specific locations. So, the team created strict "revisit consistency" trajectories: > Out-and-back: tests appearance stability
> Closed-loop: tests layout consistency
> Translation-rotation: tests identity preservation
AI
-
Strict revisit consistency trajectories test AI location memory
By
–
-
Real-time streaming inference model for interactive worlds
By
–
/5 An interactive world isn't truly interactive if it lags. This model is built specifically for real-time streaming inference. Using DMD-style distillation and an autoregressive rolling KV cache, it generates environments chunk-by-chunk from noise. When paired with asynchronous
-
DreamX-World 1.0 eliminates color drift and style mutations via long-rollout training
By
–
/6 Long-form autoregressive generation usually suffers from accumulated prediction errors, leading to color drift and style mutations. Thanks to specialized long-rollout training, DreamX-World 1.0 overcomes this limitation. It maintains stunning visual fidelity, smooth motion,
-
DreamX-World 1.0 fixes AI video environment mutation with memory persistence
By
–
/4 One major flaw in AI video is that looking away and turning back often mutates the environment. DreamX-World 1.0 fixes this with Memory-Conditioned Scene Persistence. It retrieves past frames based on camera geometry and view overlap rather than just time. By packing these
-
DreamX-World 1.0: Text to interactive video environments
By
–
Text-to-video generation has officially leveled up from passive clips to fully navigable, interactive environments.
— AlphaSignal (@AlphaSignalAI) 22 juin 2026
The DreamX Team just released a paper on DreamX-World 1.0, a general-purpose interactive world model that translates text and images into long-horizon videos you… pic.twitter.com/rVzLwhGNoKText-to-video generation has officially leveled up from passive clips to fully navigable, interactive environments. The DreamX Team just released a paper on DreamX-World 1.0, a general-purpose interactive world model that translates text and images into long-horizon videos you
-
Announcement of new GPT-5.6, Pro, and bidirectional voice models this Thursday
By
–
It looks like we are going to have a whole range of new GPT models this Thursday: GPT-5.6, 5.6 Pro, and a new bidirectional voice model. Initial tests of the voice model have been exceptional, this is exactly what I was hoping for two years ago!
-
Largest LLM-as-Judge audit shows exact-match overstates skill
By
–
The largest LLM-as-a-Judge reliability audit yet. Researchers ran 21 judges from nine providers over roughly 541,000 judgments on MT-Bench, JudgeBench, and RewardBench. Findings: Validating a judge with exact-match agreement overstates its skill, because exact match does not
-
Sakana Fugu matches Fable, Mythos with single API, Japan in big league
By
–
I don't think anyone saw this coming.
— Charly Wargnier (@DataChaz) 22 juin 2026
Sakana Fugu matches Fable and Mythos through a single model API 🤯
Japan is officially in the big league. https://t.co/DIdwrwYFdm pic.twitter.com/I9qmc9xS3II don't think anyone saw this coming. Sakana Fugu matches Fable and Mythos through a single model API Japan is officially in the big league.
-
Prompt Reinjection fixes forgetting detailed descriptions
By
–
Why do AI image generators keep forgetting your detailed descriptions? A team from Fudan University, Alibaba, and Baidu introduces Prompt Reinjection. They found that multimodal diffusion transformers gradually lose prompt info in deeper layers. Their training-free fix:
-
Pipeline for VQA data from Wikidata with LLM/VLM filters
By
–
A pipeline for creating VQA data using knowledge bases (wikidata). We demonstrated the method on world/cultural knowledge and could be extended to other domains. (there are rephrasing/filtration steps that require LLMs/VLMs, but the released data does not).