- H-FARM AI's Newsletter
- Posts
- OpenAI's "Sign in with ChatGPT" authentication
OpenAI's "Sign in with ChatGPT" authentication
PLUS: AI-designed proteins that move, Robot kickboxing tournament, AI weapons reshape Ukraine warfare & INTUITOR's self-confidence learning

In today’s agenda: | Plus, some interesting news:
|
MAIN AI UPDATES / 28th May 2025
⚙️ OpenAI's "Sign in with ChatGPT" authentication ⚙️
Third-party login system in development
OpenAI is developing a new authentication system that will allow users to sign into third-party applications using their ChatGPT accounts, similar to the widely-used "Sign in with Google" functionality. The company is actively seeking developer partners to integrate the service, which could significantly expand ChatGPT's ecosystem. This move positions OpenAI to become an identity provider in the growing AI application landscape, potentially increasing user stickiness while giving developers a streamlined way to authenticate users and access ChatGPT's capabilities.
🧬 AI-designed proteins that move 🧬
First artificial proteins that change shape
Scientists have created the world's first AI-designed proteins that can actually move and change shape like natural ones. Using artificial intelligence, researchers engineered a protein that swings to grab calcium, opening the door to custom-designed biological machines. Unlike previous static proteins, these dynamic structures could eventually lead to revolutionary applications including breaking down plastics, fighting disease, or helping crops survive climate change. This breakthrough represents a significant step toward functional biological nanotechnology that mimics natural molecular machinery.
🧠 INTUITOR: AI learns through self-confidence 🧠
Training without external feedback
Researchers from UC Berkeley and Yale have introduced INTUITOR, a revolutionary AI training method that enables language models to improve their reasoning using internal confidence signals instead of requiring correct answers or external feedback. The system measures how confident an AI feels about each word it generates, using this "gut feeling" as a learning guide. When tested on math problems, the method performed just as well as conventional training but showed superior results on programming tasks. The AIs also began showing human-like reasoning behaviors—breaking down complex problems, planning, and explaining their thinking step-by-step.
INTERESTING NEWS
🥊 Robot kickboxing tournament debuts in China 🥊
The world's first robot kickboxing tournament was recently held in China, featuring combat between two G1 robots from Unitree. The event, which quickly went viral online, showcased robots fighting in controlled MMA-style matches that tested their physical coordination, balance, and reactive AI capabilities. The tournament represents a novel intersection between entertainment and robotics research, potentially creating a new spectator sport while simultaneously advancing reactive AI and physical robotic capabilities under competitive conditions.
🔫 AI weapons transform battlefield ethics 🔫
Recent frontline reporting highlights how AI-powered weapons are reshaping the dynamics of warfare in Ukraine, challenging long-standing ethical norms and exposing gaps in international law. According to drone unit testimonies, unmanned systems now play a dominant combat role, with some Ukrainian sources estimating that drones are involved in 70–80% of battlefield casualties. The widespread use of AI-driven targeting has intensified ethical concerns, as both sides adopt tactics like disguising soldiers in civilian clothing or using human shields to exploit the limitations of autonomous systems. Experts and soldiers warn that current legal frameworks are struggling to keep pace with the accelerating use of AI in combat.

📩 Have questions or feedback? Just reply to this email , we’d love to hear from you!
🔗 Stay connected: