Can AI really reason about audio as well as it understands text or images? Researchers from CUHK, NTU, HKU, and HKUST present the first dedicated survey on audio reasoning in multimodal foundation models. The challenge: audio is continuous, time-sensitive, and packed with
RESEARCH
-
Qwen-AgentWorld: Linguistic World Models for General Agents
By
–
Qwen-AgentWorld Linguistic World Models for General Agents
-
Sakana AI CEO David Ha interviewed on TBS CROSS DIG about vision and products
By
–
Sakana AI CEO David Ha (@hardmaru) appeared on TBS CROSS DIG’s “1on1 Tech.” He talks about our founding story, latest research and technical vision, product launches including Sakana Fugu, Japan’s AI strategy, and his hopes for Japanese society. Watch the interview. https://t.co/jfRtraGpw4
— Sakana AI (@SakanaAILabs) 24 juin 2026Sakana AI CEO David Ha (
@hardmaru
) appeared on TBS CROSS DIG’s “1on1 Tech.” He talks about our founding story, latest research and technical vision, product launches including Sakana Fugu, Japan’s AI strategy, and his hopes for Japanese society. Watch the interview. -
Critique of the Real Intelligence of Claude and ChatGPT
By
–
I have no sophisticated benchmarks other than: I don't know how you can see Claude or ChatGPT conceive in the same way and write in the same way and believe that it is intelligent in any real way.
If there were 'intelligence' in there, this… -
Beginning of a major architectural change in AI development
By
–
Delighted to collaborate with @OpenRouter Products like OpenRouter Fusion and Sakana Fugu have sparked a serious conversation about dependency and resilience in AI. I believe this is only the beginning of a major architectural change to come in the development of
-
AI assists doctors in solving 18 families’ long-held medical questions
By
–
8/ The takeaway is not that AI replaces doctors. It read more than any human could, found leads, and handed them to experts who did the real medicine. For 18 families this week, a question they carried for years finally has an answer.
-
AI finds genetic leads; doctors confirm in lab
By
–
5/ It did not diagnose anyone. This matters. The model surfaced leads, candidate explanations tied to specific genes. Doctors then tested those leads in a lab and confirmed them the normal way. The AI found the thread. Humans pulled it.
-
AI model can connect old unsolved cases to new papers
By
–
4/ One detail explains why this works. A case can go unsolved, then a year later a research paper is published that finally connects a gene to a disease. No human re-reads every old case against every new paper. The model can.
-
AI reasoning model finds missed genetic link in children
By
–
3/ They gave the cases to a reasoning model, OpenAI's o3 Deep Research. The human genome has around 20,000 genes. The model read each child's symptoms, their genetic data, and the medical literature, looking for the one link a person might have missed.
-
AI solves 18 medical cases of children abandoned by doctors
By
–
Urgent: An AI just solved 18 medical cases that doctors had abandoned. All children with rare diseases that no one could identify. Here is how the AI did it: