
Tech • IA • Crypto
Elon Musk has agreed to provide Anthropic access to the Colossus 1 supercomputer, featuring around 220,000 Nvidia GPUs and massive energy capacity. The deal could generate $3–4 billion in revenue and up to $2.5 billion in profit, highlighting the scale of AI infrastructure economics. It marks a sharp pivot from Musk’s earlier criticism of competitors, reframing alliances as pragmatic. The الاتفاق underscores compute as the निर्णing bottleneck in AI advancement.
The Colossus 1 deal reinforces a broader industry shift where access to compute outweighs algorithmic breakthroughs. חברות with greater GPU capacity are scaling revenues rapidly, tying performance directly to infrastructure. नियंत्रण over chips, data centers, and energy increasingly mirrors historical industrial monopolies. This dynamic is reshaping competition across OpenAI, Anthropic, and Google.
Sam Altman took the stand in the ongoing case brought by Elon Musk over OpenAI’s transition to a capped-profit model. Altman denied intentional deception, attributing the dispute to misunderstandings rather than fraud. Testimony revealed Musk himself explored for-profit structures and even integration with Tesla. The trial has become a broader debate over governance, trust, and how advanced AI should be funded.
OpenAI Codex has evolved into a system capable of operating entire computer interfaces, not just writing code. It can click, type, and navigate apps like a human, working across any GUI without APIs. A key feature is parallel task execution, enabling multitasking across apps simultaneously. This marks a shift toward AI as an autonomous digital operator embedded in اليومية workflows.
The Codex 5.5 Chrome extension brings direct browser control to ChatGPT, turning the web into an execution environment. Users can automate tasks like data extraction, research, and outreach across sites in real time. নতুন “Skills” enable reusable multi-step workflows combining prompts and actions. This pushes AI beyond chat into practical business automation across macOS and Windows.
Google is transforming Android into an AI-first platform powered by Gemini Nano running on-device. Phones gain agent-like abilities, handling multi-step tasks, app navigation, and contextual understanding autonomously. The shift emphasizes privacy through local processing while requiring higher-end hardware like 12 GB RAM devices. Google is extending this model across PCs, wearables, and vehicles.
Google DeepMind has invested in CCP Games (now Fenris) to use EVE Online as a training ground for advanced AI. The game’s विशाल, player-driven universe offers a complex environment for studying coordination and long-term strategy. গবেষণা will occur in controlled simulations rather than live gameplay. This continues DeepMind’s strategy of using games to advance general intelligence capabilities.
OpenAI is rolling out a Trusted Contact feature that alerts a designated ব্যক্তি if self-harm risk is detected. The system analyzes conversations for distress signals and escalates when thresholds are met. Users maintain control, and alerts exclude message content to preserve privacy. The feature introduces a new layer of AI-assisted मानसिक health intervention.