Back

I've blogged on Medium after a brief break.. here is my new post: Before defence AI is tested in the real world, it needs to fail safely somewhere else. Real environments are expensive, unpredictable, and irreversible. Synthetic environments are c

"We’re not simulating for fun anymore." Every round fired in a game, every decision made under chaos - it’s all data. Data that can teach real-world AI how to think, move, and react like us. This isn’t science fiction anymore. It’s the beginning of

Harnessing GPU Power with CUDA CUDA (Compute Unified Device Architecture) is a parallel computing platform by Nvidia that unleashes the power of GPUs for more than just graphics rendering. Initially developed in 2007, CUDA enables massive parallel p

Google DeepMind just dropped AlphaEvolve‼️ A Gemini powered coding agent that evolves algorithms and outperforms AlphaTensor. It delivers: — 23% matrix speedup — 32.5% GPU kernel boost — 0.7% global compute recovery Using Gemini Flash + Pro, it re

The other night, an investor on Reddit tried to roast my startup: “Robots are better than humans… why train them to be human?” At first, it stung. But the conversation quickly forced me to clarify what Redoubt Alpha is really about. We’re not buil

NVIDIA just pledged up to $100B to OpenAI to build out multi GW AI data centers with millions of GPUs, marking one of the biggest infrastructure bets in AI history.. This isn’t just scale – it’s a signal: GPU power will be the backbone of who leads t

🗞️ Medial Bulletin 🤖 AI & Platforms • Anthropic expands Claude Enterprise capabilities, introducing stronger tool-use APIs that allow companies to connect internal databases, CRMs and software systems to AI agents. • Snowflake rolls out native ge

“Already Hooked: The Digital Funnel That Starts Before You Even Click” Most people think digital marketing begins when you see an ad—but in truth, it starts long before that. Behind the scenes, your behavior—scrolls, pauses, clicks—is being tracked

4 different ways of training LLM's. I've given a simple detailed explanation below. 1.) 𝗔𝗰𝗰𝘂𝗿𝗮𝘁𝗲 𝗗𝗮𝘁𝗮 𝗖𝘂𝗿𝗮𝘁𝗶𝗼𝗻 (𝘀𝘁𝗲𝗽-𝗯𝘆-𝘀𝘁𝗲𝗽) Prepares clean, consistent, and useful data so the model learns effectively. 1. Collect text

🔥Sarvam AI to Launch India's First Homegrown LLM🔥 🔥 Sarvam AI building India's first domestic LLM with 4,096 H100 GPUs – Bengaluru startup secured ₹99 crore in GPU subsidies under IndiaAI Mission, launching by early 2026 . 🤔 Why It Matters – Re