ERNIE 5.1 Preview just went live With a lighter, more efficient architecture, it delivers strong performance at its scale. And this is just the start — more ERNIE model updates to come at Baidu Create 2026.
LLMS
-
OpenAI Rolls Out GPT-5.5 Cyber as Anthropic Keeps Mythos Quiet
By
–
While Anthropic largely keeps its cybersecurity model Mythos under wraps, OpenAI is seizing the opportunity and making it available. How far they will roll out their GPT-5.5 cyber, and whether everyone will have access, remains to be seen. However, OpenAI is currently generating
-
Codex and GPT-5.5 Praised Across Multiple Tasks by Users
By
–
Really glad to hear Codex + GPT-5.5 has been working well for you across so many tasks! Always open to feedback on where we can make it better.
-
ChatGPT Goblin Persona Origin: Nerdy Reward Signal Analysis
By
–
ICYMI: The origin story of Goblins “the Nerdy personality was only 2.5% of ChatGPT responses but accounted for 66.7% of all “goblin” mentions. In the audit, the Nerdy reward signal preferred goblin/gremlin outputs in 76.2% of datasets.”
-
Agent-World: Scaling Real-World Environment Synthesis for General Agent Intelligence
By
–
Agent-World: Scaling Real-World Environment Synthesis for Evolving General Agent Intelligence Paper: https://
arxiv.org/abs/2604.18292
Project: https://
agent-tars-world.github.io/-/ -
Agent-World: Self-Evolving AI Agent Training Arena Using Real APIs
By
–
What if your AI agent could train itself in thousands of real-world environments instead of static datasets? Renmin University of China and ByteDance Seed introduce Agent-World, a self-evolving training arena. It automatically discovers real-world tool ecosystems (like APIs
-
Prompt obsolescence: static instructions fail with evolving AI models
By
–
But the complaint threads are missing the most important variable. Your prompts are static. The models are not. That prompt you wrote six months ago was optimized for a model that no longer exists. The instruction structure, the constraints, the output format. All calibrated
-
AI perceived as dumber, users petition for GPT-4o return
By
–
“AI is getting dumber.” That’s the most popular take on AI Twitter right now. 22,000 people signed a petition to bring back GPT-4o. Reddit threads titled “GPT-5 feels like 3.5” are hitting the front page weekly. Mainstream press picked it up. Reuters, The Guardian, Ars
-
Agentic AI Full Stack: Where Most Systems Actually Fail
By
–
Most people still think Agentic AI is just ChatGPT plus tools. They are wrong. This diagram matters because it shows the full stack. Five layers. And in my view, most failures do not happen in the models. They happen in layers four and five. 𝟭/ 𝗔𝗜 & 𝗠𝗟 𝗧𝗵𝗲