News Roundup 56

Created on March 21|Last edited on March 21
Comment
OpenAI launches new audio models with better voice and transcription tools
New gpt-4o models boost transcription accuracy and enable expressive AI voice generation
﻿https://wandb.ai/byyoung3/ml-news/reports/OpenAI-Rolls-Out-Next-Gen-Audio-Models-for-Developers--VmlldzoxMTkwNjQ1MQ﻿
Block diffusion models push text generation beyond GPT limitations
BD3-LMs blend diffusion and autoregression for faster, more coherent and scalable language modeling
﻿https://wandb.ai/byyoung3/ml-news/reports/Block-Diffusion-Language-Models-Combining-Autoregressing-and-Diffusion---VmlldzoxMTg3MjU2OQ﻿
US AISI reveals vulnerabilities in AI agents under hijacking attacks
Tests using AgentDojo show even advanced models like Claude 3.5 can be compromised
﻿https://wandb.ai/byyoung3/ml-news/reports/US-AISI-s-findings-on-agent-hijacking-evaluations---VmlldzoxMTgyODAyNw﻿
Baidu unveils ERNIE 4.5 and ERNIE X1 with reasoning and multimodal upgrades
New models focus on deep thinking, AI affordability, and native understanding of text, images, and code
﻿https://wandb.ai/byyoung3/ml-news/reports/Baidu-Unveils-ERNIE-4-5-and-ERNIE-X1--VmlldzoxMTgyODA5NA﻿
﻿
Add a comment