AI, SaaS, fintech, space-tech, and the emerging technologies reshaping India's landscape.
Stripe's recent acquisition of AI firm OpenRouter for over $8 billion is complemented by executive Nan Ransohoff's view on AI wealth driving a 'third wave' of philanthropy, drawing parallels to industrial titans.
New research from Aarjan Chaudhary, Binayak Mukherjeea, Moitreya Nandi, and Arindam Panja addresses tokenization inefficiencies for Indic languages. A regex fix cuts Nepali tokens in Llama-3/Qwen3 by 54%, while AdaptTok-Indic lowers token fertility for Bengali and Hindi, improving LLM performance and reducing inference costs.
Moonshot AI apologizes for Kimi K3.1 Pro launch issues, resetting subscriber quotas due to rate limiting. The 2.8T parameter MoE model with 1M context faced capacity strain, highlighting AI scaling challenges ahead of its full weight release by Nov 4.
Deepinder Goyal's Temple wearable opens waitlist at Rs 85,000, following multiple 'launch' announcements in 2026. Learn about its iterative public introduction.
Beijing-based Moonshot introduces Kimi K3.1 Pro, a 3-trillion-parameter-class open-weight AI model with built-in hacking curbs, addressing growing concerns about AI-powered cyberattacks. The company plans to publish the model's weights by November 4, following its Kimi K3 launch in 2024.
Honasa Consumer, parent of Mamaearth, elevates Vipul Maheshwari to Chief Product & Technology Officer, signaling a strategic commitment to agentic AI in its B2C retail operations.
More than 2,000 Akvo machines now run in at least 15 countries. Together they produce roughly 300,000 litres of drinking water a day, and none…
Rapido faces renewed scrutiny over operational integrity after a shooting incident, adding to its ongoing challenges with fragmented regulatory compliance, a CCPA fine, and unauthorized operations.
New research from arXiv details how multilingual AI models struggle with cross-script entity matching, showing a Qwen2.5-7B accuracy drop from 35.1% to 8.5% for Devanagari names. Another paper introduces a 15.45-hour Nepali-English code-switched ASR benchmark, highlighting the role of synthetic data.