Category: Anthropic

Categories
AI Anthropic Apple Google OpenAI

It’s the Harness, Stupid!

AI Harnesses

I’ve been wondering whether we’ve been looking at the AI stack from the wrong end.

Recently Kris Patel on X laid out a set of excellent questions that he’s looking to have answered as Anthropic and OpenAI move toward going public:

  1. Do you really need frontier-scale intelligence for every task?
  2. Can open-weight models provide an effective alternative to frontier models at a significant discount?
  3. Is the ultimate moat the intelligence or the harness?
  4. What other business models will the frontier labs have to adopt to make the unit economics work long term?
  5. How are you going to prevent distillation from capturing your IP and releasing it?

I’m going to explore only the third question here — moat versus harness. The other four deserve their own consideration, particularly once we have the Anthropic and OpenAI S-1s in hand, revealing for the first time the unit economics of the two largest frontier labs and how much runway they have to support the capacity they’ve contracted.

By “harness” we mean everything surrounding the model: the interface, context, memory, tools, orchestration, evaluation, permissions, and increasingly the user’s accumulated habits and data. The model supplies intelligence. The harness turns intelligence into a product.

Listening to Gavin Baker on the recent All-In episode sharpened this line of thought into something more concrete. He referenced a thought experiment from Eric Vishria: even if OpenAI or Anthropic lost their edge at the pure model layer, they would still retain significant value because of the product harness—the interface, the surrounding tooling, the orchestration—and the user familiarity and habits that have already formed around those platforms. Baker said there is a strong element of truth to it. I think he’s right, and the reasoning behind it is worth spelling out. A model can be replicated, distilled, open-weighted, or commoditized. A mature harness has network effects, switching costs, proprietary context, distribution, workflow integration, and accumulated user behavior. That’s a much harder thing to dislodge.

We are watching intelligence become more abundant and more interchangeable at the same time that the systems built around that intelligence are becoming stickier. On the developer side, the strongest examples are already clear. Cursor turns the IDE into a multi-model agentic environment with deep codebase awareness. Claude Code runs long-horizon coding agents from the terminal, planning, editing, testing, and iterating. Grok Build, Claude Cowork and similar tools emphasize parallel agents and tighter control over local context. In each case the model is a component; the surrounding system does the real work of routing, memory, tool use, and evaluation.

The consumer version of the same idea is now taking clearer shape at Apple. The rebuilt Siri AI shown at WWDC 2026 is not trying to win the pure model race. It is built as a personal harness. A system orchestrator decides what stays on-device with Apple’s Foundation Models, what moves to Private Cloud Compute, and when heavier reasoning is required. Personal context—messages, email, photos, calendar, notes, on-screen awareness—is handled largely on-device through the Spotlight semantic index and App Toolbox. Apple is designing the system so that personal context can be used without giving Apple itself access to it. Conversation history lives in a dedicated Siri app for the user to revisit. And it all syncs across all your Apple devices.

Here is the part I think matters most, and it’s easy to miss if you only read the privacy story. Apple’s Foundation Models framework doesn’t just call Apple’s own models—it’s built to support cloud models from other providers, including Claude and Gemini, conforming to a common protocol. Which means the system orchestrator, not the user, decides which model handles which task. This request goes to the on-device model. That one goes to Private Cloud Compute. A harder one might go to Claude or Gemini. The user doesn’t need to choose, and increasingly doesn’t need to know.

That’s the inversion worth exploring further. The frontier model stops being the interface and becomes a component underneath someone else’s interface. The harness chooses the intelligence. And the company that owns the harness—the OS, the identity layer, the permissions, the apps, the sensors, the notifications, the semantic index tying all of it together—has a form of leverage that has very little to do with whose model is smartest this quarter.

That reframes the subscription question too. I don’t think the right question is whether Siri gets good enough to beat ChatGPT or Claude at reasoning. I think Siri doesn’t need to win that fight at all. It needs to win a different layer entirely—the ambient assistant layer, not the reasoning layer. They’re doing different tasks. ChatGPT or Claude might remain where you go when you think, when I need to reason about something. Siri becomes where I go when I need something done: find (or make) my reservation, text my friend, find that old photograph, update my shopping list, schedule that meeting, add this thought to my notes, figure out when we’re free next week, remind me about that thing we discussed three months ago. Apple’s advantage as an ambient assistant isn’t primarily that it has your personal data. It’s that it has OS-level authority over the world that my personal data lives in.

Of course this is still early. Execution will determine how much of the architectural promise becomes daily reality. Reliability, agentic follow-through, and the quality of the on-device models will matter as much as the privacy story or the multi-model routing. But the strategic bet itself is clear, and it aligns with the broader shift: durable value is migrating toward the systems built around the models, especially systems that sit atop private, permissioned, personal context that competitors cannot easily reach. My early personal experience with the new Siri in iOS 27 betas has impressed me so far. All of this also seems to apply to Google in the context of their Pixel family of devices.

This doesn’t mean frontier labs lose. Pricing power still exists at the high end for the hardest agentic and long-horizon work. Open-weight models will continue to pressure costs and expand access. Distillation remains a real risk. But the more the capability gap narrows, and the more a harness like Apple’s can route among interchangeable frontier models rather than depend on any single one, the stronger the case that value settles into whoever controls the context—not whoever trained the model.

The coming Anthropic and OpenAI S-1s will tell us whether the frontier labs can make their economics of intelligence work. The next generation of Siri, Gemini, ChatGPT, Claude, and whatever comes after them may tell us something even more important: who gets to own the primary relationship with the user.

The model may be the engine. But the harness is where the driver sits.

What a time to be alive!

Categories
AI Anthropic Economics Stanford

Weak Links, Powerful Ideas

I’ve been thinking about bottlenecks. Not the frustrating kind you encounter in traffic or while debugging code, but the deeper structural constraints that determine how progress unfolds in our lives, organizations, and economies. A single slow step can limit an entire system, regardless of how rapidly everything else improves.

It’s an idea that feels especially relevant today. While AI capabilities continue to advance at a remarkable pace, real-world productivity gains often appear far more gradual.

Enter Chad Jones—the Stanford economist whose work has become increasingly important for anyone trying to understand AI’s long-term economic impact. This week he announced that he will join the Anthropic Institute on leave from Stanford beginning June 30.

The move is noteworthy not simply because of who Jones is, but because of the ideas he brings with him.

The Economist Who Sees Growth Through Tasks and Bottlenecks

Chad Jones (Charles I. Jones) is the STANCO 25 Professor of Economics at Stanford Graduate School of Business. He has been one of the leading scholars studying long-run economic growth: how ideas accumulate, why innovation matters, and why growth rates have remained relatively stable even as the number of researchers worldwide has expanded dramatically.

His influential work helped explain a central paradox of modern economics: adding more researchers does not automatically produce ever-faster growth because, over time, new ideas become increasingly difficult to discover.

More recently, Jones has turned his attention to artificial intelligence. Papers such as A.I. and Our Economic Future and his 2026 collaboration with Chris Tonetti, Past Automation and Future A.I.: How Weak Links Tame the Growth Explosion, examine how advances in automation may reshape economic growth in the decades ahead.

The central insight is deceptively simple:

Economic output is ultimately constrained by its weakest components.

Weak Links: The Economic Version of Amdahl’s Law

Anyone with a background in computing will recognize a familiar pattern.

Amdahl’s Law tells us that even if part of a program becomes infinitely fast, overall performance remains constrained by the portion that cannot be parallelized. Accelerating 90 percent of a workload by a factor of a million still leaves the remaining 10 percent as a hard limit on total speedup.

Jones’ “weak links” framework applies a similar logic to the broader economy.

In task-based models where tasks are complements rather than easy substitutes, every task matters. Extraordinary progress in a handful of areas does not automatically translate into extraordinary gains for the system as a whole if critical bottlenecks remain.

Historically, a large share of productivity growth has come from automation—the transfer of tasks from human labor to rapidly improving machines and capital. Jones and Tonetti argue that much of past productivity growth can be understood through this lens. The breakthrough is not merely building better machines; it is expanding the range of tasks that machines can perform.

The AI Timeline Paradox

Looking ahead, the same logic applies to AI.

Even as advanced models automate larger portions of cognitive and physical work, growth may continue to be constrained by:

  • Tasks that still require human judgment or participation
  • Regulatory and institutional frictions
  • Physical-world coordination challenges

As a result, Jones’ modeling suggests that economic growth may accelerate substantially while still unfolding more gradually than either enthusiasts or skeptics expect.

This perspective offers a useful middle ground between two popular extremes: the belief that transformative AI-driven abundance is imminent and the belief that AI’s impact will prove largely illusory. Progress can be both real and constrained. The chain remains only as strong as its weakest link.

From Theory to Practice

One reason Jones’ work resonates with me is that it extends beyond economics.

Many successful builders and leaders instinctively operate according to a weak-links philosophy. Whether in engineering, manufacturing, logistics, or organizational design, the greatest gains often come from identifying the single constraint that limits the system and focusing disproportionate effort on removing it.

Consider how Elon Musk has approached challenges at Tesla, SpaceX, and xAI. Across very different domains, a recurring pattern emerges: identify the binding constraint, concentrate resources there, remove it, and then move to the next bottleneck.

Jones’ framework provides an economic explanation for why this approach can be so effective. In systems composed of complementary tasks, relieving a key constraint can create benefits that ripple throughout the entire system.

Why This Resonates

What I find most compelling about Jones’ work is its intellectual balance.

It neither dismisses the remarkable capabilities emerging from frontier AI systems nor assumes that technological progress automatically translates into social or economic transformation. Instead, it directs attention toward the frictions, constraints, and complementarities that determine how change actually unfolds.

At a time when conversations about AI often oscillate between utopian abundance and existential catastrophe, this framework offers something rarer: a disciplined way of thinking about progress.

The weak-links perspective reminds us that the future may be shaped less by spectacular breakthroughs than by our ability to identify and address the constraints that prevent those breakthroughs from creating widespread value.

A Chain and a Compass

There is a quiet power in recognizing weak links—whether in economies, organizations, projects, or our own lives. The places where progress feels slow or frustrating are often where the greatest leverage resides.

Jones’ research provides a language for understanding those constraints, and his move to the Anthropic Institute suggests that some of the most important conversations about AI’s future may increasingly take place at the intersection of research, policy, and real-world deployment.

For that reason alone, this is a development worth watching.

If you’re interested in exploring the underlying ideas, I recommend starting with Jones’ recent papers on his Stanford faculty site, along with Anthropic’s announcement of the Institute and its mission.

Categories
AI AI: Large Language Models Anthropic

Breakout

Jack Clark doesn’t panic easily. He spent years at OpenAI watching capabilities inch upward, then left to co-found Anthropic, and has been writing his Import AI newsletter long enough to have developed — and been wrong about — many priors. So when he publishes an essay saying he has reluctantly arrived at a 60% probability that fully automated AI R&D happens by the end of 2028, the word “reluctantly” deserves some weight.

His essay, published last week and titled “Automating AI Research,” isn’t a press release or a fundraising pitch. It reads more like a man thinking out loud at the edge of something large. “I don’t know how to wrap my head around it,” he writes, which is a notable thing to say publicly when you are one of the architects of the thing you can’t wrap your head around.

The argument is built from benchmarks — not any single one, but a mosaic of them assembled to reveal a trend. SWE-Bench, the test that measures an AI’s ability to solve real GitHub issues, was at roughly 2% when it launched in late 2023. A recent Anthropic model sits at 93.9%, effectively saturating it. METR’s time-horizon plot tracks how long an AI can work independently before needing human recalibration: 30 seconds in 2022, 4 minutes in 2023, 40 minutes in 2024, 6 hours in 2025, 12 hours today. The trajectory, if it holds, suggests 100-hour autonomous work sessions by the end of this year.

Clark marshals similar progressions across AI fine-tuning, kernel design, scientific paper replication, and even alignment research itself. His throughline is the same in each: AI is now genuinely competent at the unglamorous scaffolding of AI development — the debugging, the experiment runs, the parameter sweeps, the code reviews. And crucially, it can now do these things not just faster than humans, but for longer, with less supervision.

There’s a Thomas Edison quote at the center of the essay: “Genius is 1% inspiration and 99% perspiration.” Clark’s claim is that AI has become very good at the perspiration. The question of whether it can supply the inspiration — the paradigm-shifting insight, the Move 37 — remains open. But he argues it may not need to. Most of what has moved the AI field forward has been sustained, methodical work, not lone flashes of genius. If you can automate the 99%, you have something that compounds.

There’s a data point that makes Clark’s argument feel less like forecast and more like dispatch. Last month Boris Cherny, who runs Anthropic’s Claude Code, disclosed that he hasn’t written a line of code by hand in more than two months. Every pull request — 22 one day, 27 the next — written entirely by Claude. Company-wide, roughly 70–90% of Anthropic’s code is now AI-generated. Anthropic’s stated position: “We build Claude with Claude.” The loop Clark is describing as a probability by 2028 is already running, at least partially, today.

The word Clark uses for the threshold he’s describing is not “singularity” or “AGI.” It’s quieter than that. He calls it “automated AI R&D” — the point at which a frontier model can autonomously train its own successor. It’s a specific, falsifiable thing. And he puts a number on it: 60% by end of 2028, 30% by end of 2027.

I’ve been writing about the dark software factory and the 3D printer that prints better printers, finding metaphors for what seems like an inexorable process. Clark’s essay is a different kind of writing about the same thing — the primary source document, the engineer’s log, the inventory of evidence. Reading it is a little like watching someone carefully pack boxes before a move. Each individual item seems manageable. But there are a lot of boxes.

What he’s describing — if the trend holds — is not a feature or a product launch. It’s a breakout. The moment the loop closes and the system starts building itself. He’s not certain it happens. He just thinks it’s more likely than not, and he thought you should know.

Categories
AI Anthropic Business Google

The Weight of the Bill

Technicians working among server racks with real-time performance graphs displayed on overhead screens
Technicians monitor server performance and network data in a modern data center

Jordi Visser has been making the case for months — in his weekly YouTube commentary and on his Substack — that we are living through an exponential transition that most people are measuring with the wrong instruments. I think he’s right. I found two data points this week that suggest why.

I was somewhere in the middle of an Invest Like the Best episode when Dylan Patel said it — almost as an aside, the kind of thing you drop to establish context before moving on to the point you actually came to make. His firm, SemiAnalysis, analyzes the semiconductor and AI industries for a living. And their usage of Claude, he noted, has been growing. The costs have been growing too.

Exponentially.

He moved on. I didn’t.

I think Patel’s API bill might be one of the more honest documents in the current AI moment — more honest than the analyst reports his firm produces, more honest than the earnings calls where every public company performs its AI fluency for shareholders.

Surveys bend. When you ask someone whether they’re using AI in their work, you’re asking them to self-report on a technology that has become a proxy for relevance, for not being left behind. The incentive to say yes is enormous. And even when the yes is genuine, it tells you nothing about depth — whether AI has become load-bearing in how someone actually works, or whether it’s an impressive thing they do occasionally.

Nobody pays exponentially growing API costs for show. Money is the honest witness.

What makes Patel’s situation quietly strange is the recursion in it. SemiAnalysis exists to help sophisticated investors and technologists understand this industry — and they cannot predict their own consumption curve. They are inside the exponential the same way everyone else is. They just happen to be watching their bill.

Then this morning, a different number arrived. Google announced it will invest up to $40 billion in Anthropic — $10 billion committed now, another $30 billion contingent on performance milestones. This follows a separate $5 billion from Amazon, part of a broader arrangement under which Anthropic is expected to spend up to $100 billion on compute over time.

The temptation with numbers like these is to treat them as spectacle. Forty billion dollars is so large it becomes almost aesthetic — a statement about ambition, about the kind of bets that define eras. You feel the weight of the zeros and move on.

But I keep coming back to Patel’s API bill.

Because Google’s $40 billion and SemiAnalysis’s compounding monthly costs are saying the same thing, expressed at scales so different they almost don’t seem related. One is a research firm noticing that their tool usage has quietly escaped prediction. The other is one of the most sophisticated capital allocators on earth making a bet that strains comprehension. But both are pointing at the same reality: that this technology, wherever it takes hold, does not plateau. It compounds.

We have been waiting, I think, for the moment when AI adoption becomes legibly real — some threshold event that separates the signal from the noise, the press release from the actual change. The surveys were supposed to mark that moment. The enterprise announcements. The benchmark numbers.

Patel’s aside suggests we’ve been waiting for the wrong thing. You don’t arrive at the exponential. You just eventually notice you’re already in it — in an aside on a podcast, before moving on to the point you actually came to make.

Categories
AI Anthropic Future

Escaping the Gravity of the Present

I was watching a YouTube conversation with Dario Amodei recently, and the comments he shared at the end got me thinking about how remarkably bad we all are at imagining the future.

Whenever I try to picture what the world will look like in ten or twenty years, I usually end up picturing today—just slightly shinier. If a prediction sounds too weird or disruptive, my brain automatically rejects it. It just feels too unmoored from the reality I woke up in this morning. We all have this instinct to retreat to the safety of incremental change.

But as Amodei points out, that comfort zone is exactly what blinds us. He notes that we are constantly tempted to dismiss massive shifts simply because they feel like they “can’t happen.”

“However, by extrapolating simple curves or reasoning from first principles, one often arrives at counterintuitive conclusions that surprisingly few people believe.”

It’s a strange feeling to look at a simple data curve, follow the math, and realize the logical endpoint sounds completely unhinged. The truest maps of tomorrow often look like bad science fiction to us today.

But there is a catch here, and it’s a mental trap I know I’ve fallen into before. You can’t just sit in a room and logic your way into the future. Pure logic, stripped of real-world friction, usually just leads you confidently in the wrong direction. Amodei suggests a much more grounded formula:

“The right combination of a few empirical observations and thinking from first principles can allow one to predict the future in ways that are publicly available but rarely adopted.”

This struck a chord with me. It’s easy to get swept up in purely theoretical thinking. But the better approach is to start with what is actually happening on the ground—the messy, undeniable data. From there, you strip it down to its most basic truths and follow the thread, no matter how strange the destination looks.

It takes a certain kind of intellectual courage to trust the math when your gut is screaming that things are getting too weird. But learning to decouple what is true from what feels normal might be the only real way to prepare for what is coming.

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.