🏝 ИИ-агенты начали самостоятельно играть на виртуальном острове — без заданий и наград Исследователи MIT создали виртуальный мир, в котором 13 ИИ-агентов могли свободно взаимодействовать с окружением. За 30 часов они строили конструкции из блоков, исследовали рельеф, рисовали мандалы, придумывали игры и экспериментировали с физикой. При этом агентам не ставили конкретных целей и не обещали наград. Они сами выбирали занятия, сохраняли полученные знания и использовали их в дальнейших экспериментах. Авторы подчёркивают, что исследование не доказывает наличие сознания или эмоций у ИИ. Однако оно демонстрирует, какое разнообразное поведение может возникать, когда агенту дают свободу исследовать виртуальную среду. Источник
From the original source:
Telegram — GPT Time | Нейросети | ИИ
Read full article at Telegram — GPT Time | Нейросети | ИИAI agents began to play independently on a virtual island - without tasks or rewards
🏝 AI agents began to play independently on a virtual island - without tasks or rewards MIT researchers created a virtual world in which 13 AI agents could freely interact with the environment. For 30 hours they built structures from blocks, explored the terrain, drew...
AI Summary & Key Takeaways
- 🏝 AI agents began to play independently on a virtual island - without tasks or rewards MIT researchers created a virtual world in which 13 AI agents could freely interact with the environment
- For 30 hours they built structures from blocks, explored the terrain, drew...
AI agents began to play independently on a virtual island - without tasks or rewards
Related stories
The hottest AI wearables and gadgets you should know about
We compiled a list of all the hottest AI wearables, including Muse Charm, Bee, Friend, Plaud, and more.
Walmart will allow AI agents to make purchases on their own while Amazon blocks them
Walmart will allow AI agents to make purchases on their own while Amazon blocks them The American retail chain Walmart has introduced Agent Gateway, a new system that will allow AI agents to select products and make purchases on behalf of users. Unlike the competitor...
AI news digest for October 25, 2026
Today we're highlighting a new language model from Moonshot that promises to be a leader in open source solutions, and a surprising innovation from Pennsylvania where AI-generated advertising should be
Recursive has released a model for coding agents
🌟 Recursive has released a model for coding agents The startup Recursive Superintelligence, founded by Richard Socher and emerging from the shadows in May with $650 million in investments, showed its first model - Flint. Recursive received this kind of money even before the first model, largely due to...
