Jenray

Just Do IT!

[Jenray @jenray1986]

Forget Next-Token Prediction: How Orca Builds a Latent World Model from Scratch ...... https://t.co/1qWwZ64WwN

[Jenray @jenray1986]

The End of the Docker Bottleneck: How ‘Dockerless’ is Revolutionizing AI Coding Agents ...... https://t.co/Gt4hndbniJ

[Jenray @jenray1986]

The Architecture of Individuality: Scaling PEFT for a Million Personal AI Models...... https://t.co/LiBg0Hm6og

[Jenray @jenray1986]

Accelerating LLMs with Domino: The Next Evolution of Speculative Decoding ...... https://t.co/guZhWnBsIc

[Jenray @jenray1986]

The End of Fixed-Depth Computation: Inside the Looped World Models Revolution ...... https://t.co/TWtpFgxJtU

[Jenray @jenray1986]

@dzqinxng160469 👍🏻👍👍🏼👍🏽👍🏾👍🏿

[Jenray @jenray1986]

Why Hasn’t AI Taken Our Jobs Yet? The Answer Lies in ‘Agents’ Last Exam’ ... https://t.co/SidRk4JYhu

[Jenray @jenray1986]

Beyond Generation: Why Code is the Ultimate ‘Exoskeleton’ for AI Agents https://t.co/HbZm1DlyfJ

[Jenray @jenray1986]

@dzqinxng160469 选好股票拿住，可能收益更高。

[Jenray @jenray1986]

@dzqinxng160469 厉害了👍

[Jenray @jenray1986]

@dzqinxng160469 我看好你👍，同时也看好明天修复。

[Jenray @jenray1986]

Text is the New Weight: How SkillOpt Brings Deep Learning Discipline to Prompt Optimization ...... https://t.co/ymLehpQ5N1

[Jenray @jenray1986]

The End of ‘Vibe-Based’ AI: Exposing the ‘Attribution Hallucination’ Epidemic in Multimodal LLMs ... ... https://t.co/GiOS8a0mul

[Jenray @jenray1986]

The End of the Groundhog Day Agent: How SkillOS and RL-Driven Skill Curation are Redefining Self-Evolving AI… https://t.co/hQQerzeaGj

[Jenray @jenray1986]

Beyond the Language Bottleneck: How Eywa Forges a Neural Bond Between LLMs and Scientific Foundation Models...… https://t.co/phvXCenfJ3

[Jenray @jenray1986]

From Context to Skills: How LLMs Are Teaching Themselves to Reason ...... https://t.co/pgFDbWNWY4

[Jenray @jenray1986]

The Dawn of Deployable Robot Brains: Inside the Architecture of MolmoAct2 ...... https://t.co/wQZtE2GqkS

[Jenray @jenray1986]

Beyond GRPO: How FIPO Unlocks Deep Reasoning and 10,000-Token Chains of Thought in LLMs https://t.co/0yXBC61Yz1

[Jenray @jenray1986]

Cracking the Million-Token Barrier: A Deep Dive into DeepSeek-V4’s Architecture ..... https://t.co/ForHZJqPU2