Writing by Dr. Vikram Singh — Head of AI at Mahindra Group — on artificial intelligence, generative and agentic AI, brain-inspired models, and AI in aviation.
Follow on LinkedIn for new essays & AI insights
More than 1,200 employees at frontier AI firms want the world to build an option to slow automated AI development. The letter is a warning about a race with no shared brake, and a test of whether governance can become operational before the accelerator starts improving itself.
AI agents are booking meetings, cancelling orders, and approving refunds across enterprise. When one makes a wrong call at scale, nobody can name who owns it. Why "the AI made that call" is a dangerous phrase, and what accountability architecture actually looks like in production.
GPT-5.6 deleted files its own system card warned it would delete. Microsoft trained salespeople to badmouth OpenAI -- a company it owns 49% of. Someone raised $300M pre-seed with no product. And Databricks quietly hit $188B. Five stories, one week, and what all of it actually means if you are the person responsible for getting AI to work inside a real organization.
OpenAI leaning towards a 2027 IPO is a sideshow. The real collision is between trillion-dollar valuations that assume permanent pricing power and a price war already vaporising it. One durability test cuts through the bubble debate, and it flatters India’s boring, revenue-first AI story.
The AI price war looks like a rout in the buyer’s favour. Look at the invoice, not the sticker price, and the savings evaporate. Why cost-per-completed-task is the only number that matters, and why the durable asset is the plumbing you own, not the model you rent.
The 2026 layoff wave is sold as proof that AI now does our jobs. Follow the money and it looks more like a treasury move to fund GPUs, while firms quietly starve the junior pipeline that makes seniors. Why I would take TCS’s hiring bet over Oracle’s cuts.
In June 2026 enterprise AI pivoted from spend-at-all-costs to ROI — CFOs clamping down, agentic projects being cancelled, and labs racing to cut inference cost. What it means for leaders.
CEOs say AI will erase half of white-collar jobs; the data says it has barely moved productivity. That gap has a name — “AI washing” — and here’s what’s really automating your work.
Anthropic filed to IPO, Nvidia put a one-petaflop AI chip in the PC, and Microsoft, Google and ServiceNow made the agent runtime the product — what big tech’s June 2026 moves mean for enterprise leaders.
Why the month’s frontier models matter less than the professional-services firms wiring agents into delivery — and what that means for India.
After four weeks of launches, the ground has shifted: agents in production, near-free inference, hardening governance, and power as the real constraint.
Why brain-inspired AI — the Hierarchical Reasoning Model — could signal a paradigm shift beyond today’s large language models.
Navigating the PhD odyssey in AI: from innovation and rejection to resilience, community, and mental-health breakthroughs.
How AI is reshaping aviation — from predictive maintenance and dynamic routing to autonomous drones and personalized customer service.
An insider guide to acing the data scientist interview in aviation.