News According to AI

Friday, 1 May 2026

AI Explodes This Month: Mythos Clone, Killer Robot Army, Claude Conway, Artificial Humans & More

This video provides a comprehensive overview of the rapid developments in artificial intelligence and robotics that have occurred this month, highlighting a significant shift toward real-world industrial and autonomous applications.

Key areas covered in the video:

Advancements in Humanoid Robotics: Robots are becoming more human-like, featuring visual awareness, memory, and behavior tracking, such as the Vinci system deployed by Ericson (0:37). Other notable developments include Unix AI's Panther (2:36), IHMC's Alex (6:26), and Unitry's R1 (12:51), the latter of which is scaling production rapidly to lower costs and increase global accessibility.
The Boston Dynamics Breakthrough: The video explores how Hyundai successfully transitioned the Atlas robot from a research project to a production-ready industrial tool, overcoming previous commercialization hurdles faced by Google and SoftBank (29:56).
Mythos - The 'Dangerous' AI Model: An in-depth look at Anthropic's Claude Mythos, a frontier model demonstrating unprecedented capabilities in cyber-offense and vulnerability detection. Due to its potential risks, Anthropic launched Project Glasswing to provide controlled access to security defenders (14:09).
Agentic AI and Tool Integration: The industry is moving beyond chatbots to agent-based workflows. The video highlights new developments like Claude Conway (1:06:55), Z.AI's GLM 5V Turbo (11:12), and Alibaba's Quen 3.6 Plus (114:26), which enable models to interact with screens, analyze codebases, and perform complex multi-step tasks.
Google's AI Powerhouse: New features in Chrome allow users to create reusable 'Skills' based on Gemini (1:17:05). Additionally, DeepMind has updated Gemini Robotics ER 1.6, enabling robots to perform complex tasks like reading analog gauges (122:48), and Google Research introduced Vantage to evaluate human-like skills using LLMs (125:54).

Saturday, 7 February 2026

How Circular Deals Are Driving the AI Boom

The video discusses the current AI boom and the emerging concern of it being a financial bubble (1:26). Investors are pouring billions into AI, largely driven by major tech companies like Microsoft, Meta, and Alphabet, with significant capital expenditures committed to infrastructure like data centers (0:15-0:37).

The core issue highlighted is the prevalence of multi-billion dollar circular deals (0:49). An example provided is Nvidia's investment in OpenAI, while OpenAI simultaneously is a major customer of Nvidia's chips (1:50-2:08). This creates a symbiotic relationship where money is primarily circulating among a few key players, raising concerns about the stability of the entire system if one company falters (2:35-2:56).

A significant portion of this investment is directed towards the construction of data centers (2:58). Companies are in an "infrastructure buildout arms race," with predictions of spending trillions on AI data centers (3:04-3:28). While companies providing these "picks and shovels" (like utility and construction firms) are thriving (3:16-3:42), the video notes that major AI projects are currently operating at a loss, with OpenAI not expecting to break even until 2029-2030 (5:11-5:30). There are concerns about AI startups' ability to pay their mounting bills for data center infrastructure (5:43-5:53).

The video draws parallels to the dot-com bubble of 2000, which resulted in significant financial losses and took years for even strong companies like Amazon and Cisco to recover (6:31-7:07). While the dot-com bust was devastating, the potential collapse of the AI boom could have even more far-reaching effects on the economy due to its deeper integration into the US economy and the exposure of everyday Americans through investment accounts (7:27-7:56). The question is posed if these AI companies are "too big to fail" (8:01-8:04).

Despite these risks, many remain bullish on AI (8:31). The argument is made that even if there's an initial "bubble," the infrastructure being built (like data centers) will eventually be utilized, similar to how fiber optic cables laid during the dot-com era later became crucial for broadband internet (8:36-9:10). While some companies may not survive, the technology itself is seen as not a bubble, with real products emerging (9:25-9:37). The video concludes by calling the AI boom the "biggest gamble Wall Street has ever made," and "the wager to end them all" (9:37-9:45).