Given the scale of water estimates for the "bottle of water per generated email" that come from assumptions about GPT-4's architecture, it would be greatly in the interest of @OpenAI to publish the architecture of this now-retired, three-year-old model.
GENERATIVE AI
-
Chinese AI models are now leading and frontier
By
–
That's outdated Chinese models are already leading and frontier: Alibaba, Kimi, Xiaomi, Z AI
-
Why Large Language Models Speak and Think Like Humans
By
–
Why can large language models speak and think like humans? Link: https://
github.com/hangli-hl/AI-A
rticles/blob/main/.pdf
… -
Nature article warns against using AI as student ghostwriter
By
–
Don’t let your students use #AI as a ghostwriter
by Yanjun Shen @Nature Learn more: https://
bit.ly/41QqUjd #MachineLearning #ArtificialIntelligence #EdTech -
Using Omni Text-to-Video for Sci-Fi Creative Generation
By
–
Trying to use Omni text-to-video to capture some realistic feeling scifi vibes. pic.twitter.com/ju2TteVLjD
— fofr (@fofrAI) 25 mai 2026Trying to use Omni text-to-video to capture some realistic feeling scifi vibes.
-
India Builds Foundational AI for Indian Languages
By
–
India is building its own foundational AI, trained on Indian languages, datasets, and contexts. Under the #IndiaAIMission, the IndiaAI Innovation Centre is developing multimodal models across text, speech, and vision, with deep support for Indian languages and domain-specific
-
Grok 5 Model Ready: 1.5T Parameters, Public Release in Weeks
By
–
SPEACEXAI : The next Grok model is expected to be ready for public release in 2-3 weeks. > 1.5T V9-Medium base model in comparison to 0.5T v8-Small, used for Grok 4.3
> Cursor data being used for supplementary training Grok 5? -
0.5T Model Open-Source Release Announced
By
–
We will open source the 0.5T model towards the end of this year. It should still be quite useful.
-
Local models’ progress from a year ago is amazing
By
–
We've come so far from a year ago ngl People used to tell me it's impossible for local models to be where they're now They have no idea how good this thing will be
-
EvalVerse: Expert-calibrated VLMs evaluate AI video quality
By
–
Can AI-generated video ever match professional cinema? A team from HKUST, Tencent, and Stanford introduces EvalVerse. It treats video evaluation as a science—using expert-calibrated vision-language models to judge not just if the output follows the prompt, but also its