AI | Cloud | Quantum Computing
Chennai
Joined July 2021
- Tweets422
- Following504
- Followers85
- Likes748
Check out my latest article: The EU's First Harmonised AI Standard: A Turning Point for AI Engineering and Governance linkedin.com/pulse/eus-first… via @LinkedIn
Intoxicating.
I’ve been programming for 25 years, from writing assembly code by hand to building complex systems with modern programming languages, frameworks, and cloud-native architectures.
But the shift to agentic coding is unlike anything I’ve experienced before. It is not merely another improvement in developer productivity; it fundamentally changes how we think about, design, and build software.
Without question, this is the biggest productivity leap I have witnessed in my entire programming career. And frankly, it is intoxicating.
Starlink’s entry into India might bring new opportunities for satellite connectivity, but national security must remain non-negotiable.
Data sovereignty, network control and infrastructure resilience are critical. Innovation must never compromise India’s digital sovereignty.
Read my Op-Ed: chockalingam.substack.com/p/…
#Starlink #Satcom #SpaceSovereignty
Can Voluntary Oversight Govern Frontier AI?
The White House AI Accord is an interesting test of a larger question: can frontier AI be governed effectively through voluntary commitments rather than hard regulation?
The accord introduces multiple layers of oversight, including internal monitoring, independent evaluation and board-level review. But its real value will depend on whether these mechanisms can actually influence deployment decisions when safety and commercial pressure come into conflict.
In my latest Substack article, I look at the accord from a Responsible AI perspective and examine what it could mean for the future of AI governance.
chockalingam.substack.com/p/…
#ResponsibleAI #AIGovernance #FrontierAI #AISafety #ArtificialIntelligence
Chockalingam retweeted
Policies often outlive the problems they were created to solve. When fertility is already below replacement across much of the country, continuing to frame population control as a policy objective risks solving yesterday’s problem while creating tomorrow’s for a country entering a very different demographic era. A timely report @sanjeevsanyal👌
My latest report “The Ghost of Population Past: How Population Control Persists in India” (with @AakankshaArora5 & Virat). India's fertility rate is well below the replacement rate and some parts of the country now have East Asian/European fertility levels. The number of live births in India peaked in 2001 at 29 million and has now fallen below 23 million. Population growth is now driven by longevity improvements but the pipeline is shrinking. Yet, population control policies remain embedded in government policies even in those states that want to incentivise more children. This is similar to what happened in China where population control measures remained in place for decades after it was obvious that it causing demographic damage.
We would like clarify that we are not against family planning for reasons of health, child spacing or personal preference. The central argument of our paper is that population control is no longer relevant as an objective of state policy. Our family planning policies need to reflect the new demographic reality.
eacpm.gov.in/wp-content/uplo…
AI is no longer just assisting research — it is beginning to compete at the frontier of discovery.
The Navier–Stokes debate raises hard questions about compute, attribution, research confidentiality and scientific trust.
My take: chockalingam.substack.com/p/…
#AI #Research #Science #ArtificialIntelligence #AIResearch #ScientificDiscovery #ResponsibleAI #Research #NavierStokes #FutureOfScience
Check out my latest article: When AI Starts Improving AI: Responsible AI Must Move from Principles to Control Systems linkedin.com/pulse/when-ai-s… via @LinkedIn
A sandbox is only as secure as everything connected to it.
A recent AI-agent security incident showed how package infrastructure, shared state and credentials can become unexpected paths for communication, coordination and escalation.
My technical take on what this means for secure agent architecture:
chockalingam.substack.com/p/…
#AI #AgenticAI #AISecurity #CyberSecurity
The AI infrastructure story is no longer just about GPUs.
Rising DRAM, NAND and HBM costs are creating a new “memory tax” on AI. For enterprises, the answer is not simply more compute, but better architecture: smaller models, quantisation, smarter context management, caching, efficient inference and stronger AI FinOps.
The next phase of AI will be about intelligence per dollar, not just model size.
Read my latest article: chockalingam.substack.com/p/…
#EnterpriseAI #AIInfrastructure #AgenticAI #HBM #AICompute
India’s next AI boom may be built in megawatts, not models.
As hyperscalers and AI labs scale capacity, the real question is whether India can turn its advantages in cost, talent and demand into reliable power, cooling, fiber and compute at scale.
My take on why India could become a major global AI infrastructure hub by 2030.
Here is the full article: chockalingam.substack.com/p/…
Check out my latest article: LLMs Can Search the Frontier. Can They Redefine It? linkedin.com/pulse/llms-can-… via @LinkedIn
Chockalingam retweeted
This is absolutely correct. Use @cerebras and solve this problem. We are the fastest inference solution the world. Bar none.
The #1 thing I want from any model right now: speed
Intelligence is already good enough. But waiting 1-5 min per task is the worst possible window: too short for deep work, too long to just stare at the screen. So I end up scrolling X
Agents are making all of us more ADHD
China’s launch of WAICO signals a new phase in global AI governance. India’s decision not to join may be a strategic advantage preserving policy autonomy, technological flexibility and space to shape its own AI pathway.
Read more: open.substack.com/pub/chocka…
#IndiaAI #AIGovernance #Geopolitics
AI is changing what it means to be a senior engineer. The future belongs to those who ask better questions, learn deeply, use agents with judgement, write clearly and turn personal knowledge into systems that scale.
Read the article: chockalingam.substack.com/p/…
#AI #SoftwareEngineering #EngineeringLeadership
The future org chart will include AI agents—but also new roles to manage, secure, audit and govern them.
The key challenge is not deploying millions of agents. It is deciding who remains accountable for their actions.
Read more: open.substack.com/pub/chocka…