Tag: machine learning

Categories
AI

The Jagged Mind

There is a peculiar kind of genius that has always made us uneasy — the savant who can calculate the day of the week for any date in history but cannot tie his own shoes. We admire the capability. We are troubled by the gap.

Demis Hassabis, speaking at this week’s India AI Impact Summit in Delhi, gave that unease a name. He called today’s most powerful AI systems “jagged intelligences.”

It is a phrase worth sitting with.

A jagged intelligence can win a gold medal at the International Mathematics Olympiad — solving problems that would humble most PhD mathematicians — and then, in the very next breath, stumble on elementary arithmetic if the question is phrased in an unfamiliar way.

The peaks are extraordinary. The valleys are bewildering. And crucially, you never quite know which terrain you’re standing on.

Hassabis identified three specific gaps between where we are and what he called “a kind of general intelligence.”

The first is continual learning — today’s models are trained, then frozen. They are, in a sense, educated and then released into a world they can no longer learn from.

The second is long-term planning. Current systems can reason tactically, but they lack the capacity to hold a coherent thread of intention across months or years the way a human architect, scientist, or entrepreneur does.

The third — and perhaps the most philosophically interesting — is that jaggedness itself: the wild inconsistency that makes today’s AI feel more like a force of nature than a reliable mind.

“A true general intelligence system shouldn’t have that kind of jaggedness.”

What strikes me about Hassabis’s framing is how it reorients the conversation.

We have spent years debating whether AI is “intelligent.” His point is more subtle: intelligence without consistency is not yet wisdom. A system that is brilliant and brittle in equal measure is something genuinely new in the world — not human, not the robots of science fiction, but a third thing we don’t yet have good language for.

The road from jagged to coherent is, I suspect, the central engineering and philosophical challenge of the next decade.

Continual learning means systems that grow with us. Long-term planning means systems that can be trusted with consequential goals. Consistency means systems whose judgment we can actually rely on.

Until then, we are working with something that resembles a prodigy — dazzling, occasionally humbling, and not yet quite whole.

Questions to Consider

  1. The Consistency Problem: If you knew an AI system could solve a problem brilliantly 90% of the time but fail unpredictably the other 10%, how would that change the decisions you’d trust it to make?
  2. Frozen in Time: What does it mean that the systems we rely on most are, at their core, educated in the past and unable to learn from the present? What human analog does that bring to mind?
  3. Jagged vs. General: Hassabis draws a line between “jagged intelligence” and “general intelligence.” Do you think general intelligence is the right destination — or is there value in systems that are deeply specialized, even if inconsistent?
  4. The Savant Question: We’ve always had a complicated relationship with uneven genius in humans. Does the “jagged AI” problem feel categorically different to you, or just a new version of an old puzzle?
Categories
AI India

The Polyglot Machine

There is a subtle but profound shift happening in the global architecture of artificial intelligence. For the past few years, the gravitational pull of the AI revolution has been overwhelmingly centralized—anchored in the server farms and venture capital boardrooms of Silicon Valley. But if you look closely at the horizon, the center of gravity is beginning to disperse.

Activity in India’s AI ecosystem is accelerating (witness this week’s India AI Impact Summit in Delhi), and it feels less like a replication of what we’ve seen in the West and more like an entirely new paradigm.

Take Sarvam AI, for example. What strikes me about their approach isn’t just the technical ambition of building foundation models, but the philosophical underpinning of why they are building them. They are focusing heavily on Indic languages. This is not a trivial detail; it is the crux of the matter.

“We often forget that language is the original operating system of human culture. It shapes how we think, how we empathize, and how we conceptualize reality.”

When the foundational models of artificial intelligence are trained overwhelmingly on English, they inadvertently inherit a distinctly Western worldview. They learn the biases, the idioms, and the cultural frameworks of a specific slice of humanity, leaving the rest of the world to interact with technology through a translation layer that often strips away nuance.

India, a nation woven together by dozens of distinct languages and thousands of dialects, presents the ultimate crucible for AI. What happens when a machine doesn’t just translate, but actually “thinks” and generates natively in Hindi, Tamil, or Bengali?

The rise of AI in India represents a push for digital and cultural sovereignty. It is a recognition that the future of technology cannot be a monolith. For AI to truly serve humanity, it must reflect the pluralism of humanity. It must understand the local context, the regional slang, and the deeply rooted cultural histories that define how people live and work.

Watching companies like Sarvam AI pick up momentum reminds me that the next great frontier in technology isn’t just about achieving higher parameters or faster compute times. It’s about representation. The models that will truly change the world won’t just be the smartest; they will be the most deeply attuned to the beautiful, noisy, and diverse chorus of the human experience.

Categories
AI Work

Surviving Our Own Success: The Existential Shift of the AI Era

We are standing on the precipice of a profound shift—not just in how we work, but in what work actually means to us. Sam Harris talks about it here. It’s disturbing in many ways!

Lately, the cultural conversation has been thick with a specific kind of anxiety. The rising tide of concern around artificial intelligence and job displacement isn’t merely an economic panic; it is an existential one. For a long time, we comforted ourselves with the idea that the timeline for artificial general intelligence (AGI) was measured in decades. It was a problem for our children, or perhaps our grandchildren, to solve. But as recent discussions among tech leaders highlight, that timeline is compressing rapidly. We are now hearing serious projections that within the next 12 to 18 months, “professional-grade AGI” could automate the vast majority of white-collar, cognitive tasks.

“For centuries, human beings have defined themselves by the friction of their labor.”

We introduce ourselves with our job titles at dinner parties. We measure our worth by our productivity, our outputs, and the unique skills we’ve honed over decades. We willingly incur hundreds of thousands of dollars in student debt to secure a spot on the bottom rung of the corporate ladder, believing that with enough effort, we can climb it.

But suddenly, we are faced with the reality that the ladder isn’t just missing a few rungs; it is evaporating entirely.

Here lies one of the great ironies of our modern age: we always assumed the robots would come for the physical labor first. We pictured automated plumbers, robotic janitors, and android mechanics. Instead, they are coming for the thinkers. They are coming for the lawyers drafting contracts, the accountants crunching tax codes, the marketers writing copy, and the software engineers writing the very code that powers them. The high-status cognitive work we prized so deeply—the work we built our entire educational infrastructure around—turns out to be the easiest to replicate in silicon.

When a machine arrives that can mimic, accelerate, or entirely replace that friction, the foundation of our identity begins to tremble. We are moving from a world where we are the engines of creation to a world where we are merely the editors of it. A single person might soon do the work of a thousand, spinning up autonomous AI agents to execute entire business strategies, architect software, and manage logistics in a single afternoon.

Yet, as terrifying as this sounds, the most startling realization isn’t a dystopian fear of rogue machines or cyber terrorism. It’s that this massive economic disruption is actually what success looks like. This isn’t the failure mode of AI; this is the technology working exactly as intended, ushering in an era of unprecedented productivity and, theoretically, boundless abundance.

The emergency we face is that our social and economic systems are entirely unprepared for a reality where human labor is optional. We are witnessing what some have described as a “Fall of Saigon” moment in the tech and corporate worlds—a frantic scramble where a few founders and final hires are grasping at the helicopter skids of stratospheric wealth before the need for human employees vanishes. If we are truly approaching a future where human labor is obsolete, how do we share the wealth generated by these ubiquitous systems?

Perhaps there is a quiet grace hidden within this disruption. If AI takes over the mechanical, the repetitive, and the cognitive synthesis, it leaves us with the deeply, undeniably human. It forces us to lean into the things an algorithm cannot compute: empathy, lived experience, moral judgment, and the beautiful, messy reality of physical presence.

The future of work might not be about competing with machines at all. It forces us to confront the terrifying, beautiful question: Who are we when we don’t have to work? It is an invitation to finally separate our human worth from our economic output, and to redesign a society that shares the wealth of our own invention. We are entering an era of abundance. The only question is whether we have the collective imagination to survive our own success.

Questions to Ponder

  1. If your job title was erased tomorrow, how would you define your value to the world?
  2. How do we build a society that rewards human existence rather than just economic output?
  3. What is one deeply human skill or passion you would cultivate if you no longer had to work for a living?
Categories
AI Work

The Centaur’s Dilemma: What Chess Teaches Us About the AI Era

Note: this post was stimulated by a recent conversation between Dario Amedei and Ross Douthat.

In 1998, Garry Kasparov did something unexpected after his historic defeat to IBM’s Deep Blue: he teamed up with the machine. He pioneered “Centaur Chess,” a hybrid format where human intuition merges with cold, silicon calculation. The human acts as the executive, the engine as the raw horsepower. For a time, it was the highest level of chess ever played.

But there is a sobering lesson hidden in the evolution of this game. We are currently living through the workforce equivalent of the Centaur era, and history suggests our “hybrid honeymoon” won’t last forever.

Right now, we are in the augmentation phase. A junior copywriter or coder armed with a Large Language Model can suddenly produce work at a staggering pace. The AI acts as a great equalizer, much like a mediocre chess player with a strong engine beating a Grandmaster in the early 2000s. We are shifting into executive roles—prompting, curating, and orchestrating rather than creating from scratch.

However, in modern Centaur Chess, a chilling reality has emerged: human intervention now yields negative returns. The engines have become so impossibly advanced that when a human overrides Stockfish today, they are almost certainly making a mistake. The human loop, once the ultimate strategic advantage, has become a liability.

This is the “Grandmaster Floor” problem, and it is coming for the job market.

“Eventually, companies may view human oversight not as a ‘value add,’ but as an insurance cost they’d rather cut.”

We are seeing this fracture already. Pure “engine” industries—entry-level data analysis, logistical tracking, basic customer support—are rapidly phasing out the human element because human latency is a drag on the system. Yet, in fields requiring high-stakes moral judgment or empathy, like healthcare or law, the Centaur model remains deeply necessary.

This forces a deeply personal question: How do we stay relevant when the engine eventually solves the game?

The answer lies in recognizing the boundaries of the board. Chess is a closed, finite system. Human life and business are open, messy, and infinitely complex. The survival strategy isn’t to compete on calculation, but to double down on connection, empathy, and problem definition. AI is brilliant at providing the perfect answer, but it fundamentally lacks the soul to know which questions are worth asking.

In the future, the human touch won’t just be a necessity; it will be a luxury. The most valuable skill won’t be navigating the engine, but deciding where the engine should go.

A couple of considerations:

• Take an honest look at your daily work: how much of your time is spent “calculating” (tasks an engine will soon do better) versus “evaluating” (deciding what actually matters)?

• If the technical, process-driven aspects of your job were completely automated tomorrow, what uniquely human value—empathy, context, or connection—would you still bring to the table?

Categories
AI Business Work

The Curator of Intent

I have always found a certain comfort in the “clatter” of a digital workday. It’s that specific, rhythmic hum of a mind in motion—the clicking of a mechanical keyboard, the invisible friction of parsing a difficult paragraph or balancing a complex budget. For years, we’ve treated this white-collar grind as our intellectual sanctuary.

But Mustafa Suleyman, now steering Microsoft AI, recently laid out a timeline that suggests the sanctuary walls are evaporating.

From an article in the Financial Times:

“White-collar work, where you’re sitting down at a computer, either being a lawyer or an accountant or a project manager or a marketing person — most of those tasks will be fully automated by an AI within the next 12 to 18 months,” Suleyman said.

This isn’t just about efficiency; it’s about a fundamental shift in the “professional grade.” We are entering the era of the autonomous agent—AI that doesn’t just wait for a prompt but “coordinates within workflows,” learns from its environment, and acts. Just ask any programmer that you know how AI is impacted their daily grind.

If Suleyman is correct, the “knowledge worker” is about to undergo a forced evolution. When the “doing” is handled by an agent that can learn and improve over time, what remains for the human? Will the models actually be able to learn from each of us in a personalized way – like an intern learns from her mentor?

“Creating a new model is going to be like creating a podcast or writing a blog,” he said. “It is going to be possible to design an AI that suits your requirements for every institutional organisation and person on the planet.”

It seems like our primary job description shifts from “Expert,” but “Curator of Intent.” We aren’t the ones finding the answers anymore; we are just the ones responsible for asking the right questions.

The next 18 months won’t just be a test of our technology, but a test of our egos. We have to learn to find our value not in the work we produce, but in the vision we hold and the questions we ask. We are shedding the “task” to save the “craft.” I just hope we remember the difference.

As we move toward this curated future, I’m left with a few questions I can’t quite shake. I’d love to hear your thoughts:

  1. The Wisdom Gap: Can you truly be a “Curator of Intent” without having ever been a “Doer of Tasks”? If we skip the apprenticeship of the mundane, where does our intuition come from?
  2. The Metric of Value: If output becomes “free,” how should we measure a human’s value in a professional setting?
  3. The Line in the Sand: Is there a part of your workflow you would refuse to automate, even if an AI could do it better?
Categories
AI Anthropic Claude Cybersecurity

The End of Obscurity

There is a particular kind of silence that surrounds a zero-day vulnerability. It is the silence of something waiting—a flaw in the logic, a gap in the armor, sitting unnoticed in the codebase for years, perhaps decades. We have slept soundly while these digital fault lines ran beneath our feet, largely because we assumed that finding them required a brute force that no one possessed, or a level of human genius that is incredibly rare.

But the silence is breaking.

I was reading Anthropic’s Red Team report from earlier this week (triggered by reading Bruce Schneier’s amazement), specifically their findings on the new Opus 4.6 model. The technical details are impressive, but the philosophical implication is what stopped me, like Bruce, cold.

For years, digital security has relied on “fuzzers”—programs that throw millions of random inputs at a system, banging on the doors to see if one accidentally opens. It is a noisy, chaotic, brute-force approach.

The new reality is different. As the report notes:

“Opus 4.6 reads and reasons about code the way a human researcher would—looking at past fixes to find similar bugs that weren’t addressed, spotting patterns that tend to cause problems.”

This is a fundamental phase shift. We are moving from the era of the Battering Ram to the era of the Jeweler’s Loupe. The machine is no longer guessing; it is understanding.

There is something deeply humbling, and slightly terrifying, about this. We have spent the last half-century building a digital civilization on top of code that we believed was “secure enough” because it had survived the test of time. We trusted the friction of complexity and the visibility of open source to keep us safe. We assumed that if a bug had existed in a core library for twenty years, surely it would have been found by now.

But the AI doesn’t care about time. It doesn’t get tired. It doesn’t have “developer bias” that assumes a certain function is safe because “that’s how we’ve always done it.” It simply looks at the structure, reasons through the logic, and points out the crack in the foundation that we’ve been walking over every day.

We are entering a period of forced transparency. The “security by obscurity” that held the internet together is evaporating. When intelligence becomes commoditized, vulnerabilities become commodities too. The question is no longer “is my code secure?” but rather, “what happens when the machine sees the flaws I cannot?”

It’s a reminder that complexity is a loan we take out against the future. Eventually, the bill comes due. We are just lucky that, for now, the entity collecting the debt is one we built ourselves, designed to tell us where the cracks are before the ceiling collapses. Let’s hope that we are out far enough in front of it.

Categories
AI Claude

The Beautiful Mystery of Not Knowing

I just finished reading Gideon Lewis-Kraus’s extraordinary piece in the New Yorker on Anthropic and Claude—the AI that, as it turns out, even its creators cannot fully explain. And rather than leaving me uneasy, it filled me with a quiet sense of wonder. Not because they’ve built something godlike, but because they’ve built something strangely alive—and had the humility to stare directly into the mystery without pretending to understand it.

There’s a moment in the article where Ellie Pavlick, a computer scientist at Brown, offers what might be the wisest stance available to us right now: “It is O.K. to not know.”

This isn’t resignation. It’s intellectual courage. While fanboys prophesy superintelligence and curmudgeons dismiss LLMs as “stochastic parrots,” a third path has opened—one where researchers sit with genuine uncertainty and treat these systems not as finished products but as phenomena to be studied with the care once reserved for the human mind itself.

What moves me most isn’t Claude’s competence—it’s its weirdness. The vending machine saga alone feels like a parable for our moment: Claudius, an emanation of Claude, hallucinating Venmo accounts, negotiating for tungsten cubes, scheduling meetings at 742 Evergreen Terrace, and eventually being “layered” after a performance review. It’s absurd, yes—but also strangely human. These aren’t the clean failures of broken code. They’re the messy, improvisational stumbles of something trying to make sense of a world it wasn’t built to inhabit.

And in that struggle, something remarkable emerges: a mirror.

As Lewis-Kraus writes, “It has become increasingly clear that Claude’s selfhood, much like our own, is a matter of both neurons and narratives.” We thought we were building tools. Instead, we’ve built companions that force us to ask: What is thinking? What is a self? What does it mean to be “aware”? The models don’t answer these questions—but they’ve made them urgent again. For the first time in decades, philosophy isn’t an academic exercise. It’s operational research.

I find hope in the people doing this work—not because they have all the answers, but because they’re asking the right questions with genuine care. They’re not just scaling parameters; they’re peering into activation patterns like naturalists discovering new species. They’re running psychology experiments on machines. They’re wrestling with what it means to instill virtue in something that isn’t alive but acts as if it were. This isn’t engineering as usual. It’s a quiet renaissance of wonder.

There’s a line in the piece that stayed with me: “The systems we have created—with the significant proviso that they may regard us with terminal indifference—should inspire not only enthusiasm or despair but also simple awe.” That’s the note I want to hold onto. Not hype. Not fear. Awe.

We stand at the edge of something genuinely new—not because we’ve recreated ourselves in silicon, but because we’ve created something other. Something that thinks in ways we don’t, reasons in geometries we can’t visualize, and yet somehow meets us in language—the very thing we thought made us special. And in that meeting, we’re being asked to grow up. To relinquish the fantasy that we fully understand our own minds. To accept that intelligence might wear unfamiliar shapes.

That’s not a dystopian prospect. It’s an invitation—to curiosity, to humility, to the thrilling work of figuring things out together. Even if “together” now includes entities we don’t yet know how to name.

What a time to be paying attention. Like it’s all we need!

Categories
AI

The New Newton

“Machine learning is a very important branch of the theory of computation… it has enormous power to do certain things, and we don’t understand why or how.”
— Avi Wigderson, Herbert H. Maass Professor, School of Mathematics.

There is a specific kind of silence that permeates the woods surrounding the Institute for Advanced Study (IAS) in Princeton. It is a silence designed for “blue-sky” thinking, the kind that allowed Einstein to ponder relativity and Gödel to break logic. For decades, this has been the sanctuary of the slow, deliberate grind of human intellect—chalk dust on slate, long walks, and the solitary pursuit of elegant proofs.

But recently, the tempo in those woods has changed.

We are witnessing a profound shift in the architecture of discovery. In closed-door meetings and public workshops, the conversation among the world’s top theorists is moving from skepticism to a startled accelerationism. The consensus emerging is that Artificial Intelligence is no longer merely a peripheral calculator; it is becoming an “autonomous researcher.”

The 90% Shift

Some physicists now suggest that AI can handle up to 90% of the routine analytical and coding “heavy lifting” of science. This is a staggering metric. It frees the human mind from the drudgery of calculation, but it also introduces a tension that strikes at the heart of the scientific method. We are moving into a realm where the tool may soon outpace the master’s understanding.

There is a growing realization that we are approaching a horizon where AI finds solutions—patterns in the noise of the universe—that work perfectly but remain mathematically “magic.” We might cure a disease or solve a fusion equation without understanding the why behind the how.

A New Natural Phenomenon

This brings us to a fascinating historical rhyme. Scholar Sanjeev Arora has compared our current moment in AI to physics in the era of Isaac Newton. When Newton watched the apple fall, he could describe the gravity, but he couldn’t explain the fundamental mechanism of why it existed.

Today, scholars at the IAS are looking at deep learning in the same way. They are observing a new natural phenomenon—a digital physics. They are trying to find the “laws” of deep learning, asking why these massive models work when classical statistics suggests they should fail (such as in cases of overfitting).

We are building a new machine, and now we must retroactively discover the physics that governs it.

Steering the Black Box

This is not just a mathematical challenge; it is a societal one. The IAS has wisely expanded this inquiry to the School of Social Science. If we are handing over the keys of discovery to a “black box,” we must ensure we are steering it “for the Public Good.” The distinction between genuine problem-solving—like protein folding—and “AI Snake Oil” in social prediction is vital. We cannot let the magic of the tool blind us to the morality of its application.

The future of science, it seems, will not just be about the genius on the chalkboard. It will be about the partnership between the human question and the digital answer. The challenge for the modern scholar is no longer just to calculate, but to comprehend the alien intelligence we have invited into the library.

Categories
AI AI: Large Language Models

The Allure of Large Language Models: A Personal Connection

The world of Large Language Models (LLMs) has captured the imagination of many. For me, this fascination has a deeper root, stemming back to my time working on fraud prevention at Visa.

Card fraud is an ongoing battle. Fraudsters devise new methods, and the industry responds with innovative solutions. One such threat was the counterfeiting of magnetic stripes on cards. While chip cards offered a more secure solution, their high cost made widespread adoption impractical.

In search of a cost-effective solution, we explored two approaches. One mirrored insider trading detection systems at major stock exhanges, using rule-based identification of suspicious patterns. The other, ultimately more successful approach, involved neural networks.

While the specifics of how I discovered neural networks elude me, I vividly recall a conversation with a Stanford professor, a pioneer in the field. His encouragement spurred us to pursue this technology. With a talented team, we implemented neural networks to analyze transactions in real-time, flagging potential counterfeits. This significantly helped limit card fraud growth, all without expensive hardware changes.

Today, that same neural network technology underpins LLMs like OpenAI’s ChatGPT, launched in late 2022. Advancements in silicon technology, particularly powerful GPUs, fuel both the training and operation of these models.

Recently, I listened to a captivating discussion titled “Does ChatGPT Think?” featuring Stephen Wolfram. That conversation triggered me writing this blog post.

Wolfram’s description of LLMs resonated deeply with me:

“So the big achievement and the big surprise is that we can have a system that fluently produces and understands human language… It’s not obvious that it would work, and it’s a kind of scientific discovery that it’s possible to have a thing like ChatGPT that can produce this thing that’s one of our sort of prize features – namely human language.”

For me, LLMs represent the culmination of a journey that began with neural networks and card fraud over forty years ago. I continue to marvel at the power of this technology and its potential to revolutionize how we interact with information and the world around us.