AGI is Already Here: The 13 Most Hardcore AI Conversations from Sequoia Capital's Annual Meeting

Introduction

At the end of April 2026, Sequoia Capital held its fourth annual AI Ascent conference in San Francisco. The event brought together core AI industry players like OpenAI, DeepMind, Anthropic, NVIDIA, and Waymo, as well as startups betting on emerging directions like ElevenLabs, XBOW, Recursive Intelligence, and Starcloud. The 13 dialogues spanned foundation models, programming paradigms, robotics, autonomous driving, chip design, space computing, and novel computing architectures, covering essentially the most cutting-edge frontiers of the current AI industry.

Compared to previous years, the tone of this AI Ascent was more direct: AI is no longer just a tool for improving efficiency but is beginning to enter real workflows and take over complex tasks previously only done by humans. In their opening speech, Sequoia called this the arrival of "functional AGI" – not that machines are already equivalent to humans in every dimension, but from a commercial and productivity perspective, long-horizon agents have crossed the threshold from demonstration to practical use.

This is the core backdrop of the conference: when intelligence becomes cheap, callable, and scalable, the focus of AI competition is shifting from "can the model do it" to "how to connect it to the real world." Software, services, organizations, hardware, energy, security, and physical space may all be redesigned as a result.

The story Sequoia attempted to tell is clear: intelligence is no longer a luxury good but is becoming a new industrial raw material. What truly matters in the next phase might not be who has the smarter model, but who can understand customers faster, reorganize processes, orchestrate agents, and convert this cheap intelligence into sustainable business systems.

Therefore, this conference discussed not just the next steps for AI technology, but a larger question: when machines can shoulder more and more intellectual labor, how should humans, companies, and society redefine their own value.

Key Threads Throughout the Conference

First, intelligence is becoming a commodity.

Sequoia likened this transformation to "aluminum" in the late 19th century: once more expensive than gold, it became a readily available industrial material within decades due to the popularization of electrolysis. Today, PhD-level expertise and cognitive barriers that once defined middle-class competitiveness might be undergoing a similar fate. High-level intelligence is no longer naturally scarce; instead, it is being mass-produced, called upon, and distributed by models.

Second, the bottleneck is shifting from machine to human.

Greg Brockman uttered a phrase that was repeatedly quoted throughout the conference: when agents can work autonomously, human attention will become the scarcest resource in the entire economy. Karpathy expressed the same judgment more bluntly: when machines can handle almost all execution details, the only ability humans cannot afford to lose is understanding what they truly want. The problem is no longer whether the machine can do it, but whether humans can set the right goals, judge if the results are reliable, and decide what is worth accomplishing.

Third, programming is being solved, but organization is not.

Internally at Anthropic, a large amount of code is now generated by models, with different agents even collaborating autonomously on Slack. Boris Cherny's assessment goes further: the real moat is no longer a particular model version but the degree to which an organization's structure is "AI-native." For existing companies, this is an unfriendly conclusion – because the gap doesn't just come from tool proficiency, but from whether the company is willing to redesign processes, permissions, collaboration methods, and management structures around agents.

Fourth, AI is returning from the digital world to the physical world.

Jim Fan's robotics, Waymo's 20 million autonomous rides, and ElevenLabs' emotional speech all demonstrate from different angles that AI is no longer just a screen-based tool for processing text, code, and images. It is beginning to understand and intervene in light, sound, force, motion, and space. In the past decade, "software is eating the world" was the main thread; next, AI may directly enter the physical world, transforming cars, factories, robots, voice interaction, and physical manufacturing itself.

Fifth, the endgame of computing power lies in the physical fundamentals.

As land, electricity, and cooling for data centers hit their limits, a group of more radical companies offer different solutions: Starcloud wants to send chips into space, Recursive lets AI design chips autonomously, Unconventional AI attempts to bypass the von Neumann architecture to mimic the brain, and Flapping Airplanes directly questions "brute-force scaling" itself – if humans can learn the same skills with far less data, then today's AI algorithms might be fundamentally too inefficient from the ground up. The endpoint of the computing power race is moving from buying more GPUs to deep structural reform in energy, chips, architecture, and data efficiency.

Sixth, security has entered the asymmetric battlefield of "AI vs AI."

XBOW's agent topping the global white-hat hacker leaderboard means AI is no longer just an auxiliary tool for security researchers but an autonomous attack system capable of independently discovering, verifying, and exploiting vulnerabilities. More concerning, as the capabilities of open-source models improve, this kind of attack capability could spread rapidly within the next 6 to 9 months. Cybersecurity is no longer a confrontation between human hackers but an AI arms race with a countdown already started.

Putting these threads together, it's clear that the AI industry in 2026 is in an uncomfortable position: technological capabilities have already run far ahead of product forms, organizational structures, and social rules. Models are getting stronger every day, but the "containers" that receive them – whether enterprise processes, application interfaces, or human attention itself – have not kept up.

The entire conference discussion essentially revolved around answering the same question: in a world where machines can perform more and more intellectual labor, what is left for humans?

Sequoia's answer is somewhat counterintuitive: emotion, trust – things that cannot be mass-produced. Brockman's answer is "what you want," Karpathy's answer is "whether you can judge if the machine is doing it right." These answers ultimately point to the same thing: when intelligence itself is no longer scarce, intent, judgment, and relationships become the new hard currency.

Here is a summary of all 13 dialogues from the conference.

Forum Summaries

Keynote Speeches

Sequoia Partners Opening Keynote: This Is AGI

The speakers Pat Grady, Sonya Huang, and Konstantine Buhler are three core partners on the AI investment team at Sequoia Capital. Sonya Huang was the author of the globally viral 2022 article "Generative AI: A Creative New World," considered one of the earliest institutional investors to systematically bet on generative AI. The three co-authored the 2026 essay "This is AGI," which served as the intellectual framework for this conference. Sequoia Capital itself is Silicon Valley's most historic top-tier venture capital firm, with early investments in Apple, Google, Nvidia, Stripe, and OpenAI.

AI is a "Computing Revolution" that completely overturns the nature of information processing, not merely a "Communications Revolution" that accelerates distribution. Previous internet and mobile eras only changed the transmission path of information, but AI changes the underlying logic of information generation, causing the floor (technical foundation) on which developers build applications to shift daily. The significance of this judgment lies in the fact that during a "rainstorm of instability," the traditional stable tech stack is a thing of the past; developers must learn to dance with the ever-evolving model foundation.

AI will tap into a $10 trillion market, ten times larger than traditional software, by directly delivering "professional services." The TAM (Total Addressable Market) for global software is only a few hundred billion dollars, whereas the US legal services market alone is $400 billion, equivalent in size to the entire software industry. This advocates a key transformation: AI's commercial value is no longer about selling tools to humans but directly using agents to take over and deliver high-value work originally done by human experts.

From a business practice perspective, long-duration agents capable of autonomously handling failure signify that AGI (Artificial General Intelligence) has arrived. If a system can be dispatched to perform a task, self-correct after failure, and persist to the end, it is functionally equivalent to AGI. This counterintuitive judgment reminds us: stop obsessing over academic definitions; an AI capable of independent execution has evolved from a "faster horse" to a "car" that changes the dimension of competition, achieving a 10 to 40-fold leap in efficiency.

In times of rapidly shifting foundational capabilities, the only logic for building a moat is to be "extremely close to the customer." The MAD strategy – Moats, Affordance (referring to how intuitive and easy-to-use a product is), and Diffusion – advocates locking in value using a customer-back approach rather than a tech-out approach. Since human needs change much slower than model capabilities, deep integration with the customer is more sustainable than chasing models.

Agent autonomy is jumping from "minute-level assistants" to "hour-level autonomous employees." The meter chart (task persistence metric), measuring how long a model can stay on track during complex tasks, has leaped from minutes a year ago to several hours now, sufficient to support dark factories (fully autonomous business processes) that require no human review. This means the productivity bottleneck has been broken, and extraordinary iterations like "rewriting 8 million lines of code in 6 weeks" are becoming the norm.

Human society is on the eve of a "Cognitive Industrial Revolution," where machines will handle 99.9% of global mental labor. Just as the Industrial Revolution replaced 99% of physical labor with engines, the vast majority of future analysis, decision-making, and creation will be undertaken by neural networks. The implication of this judgment is that intelligence will no longer be a monopolized resource for humans but a low-cost, industrial-grade consumable that can be mass-produced and called upon on demand.

Advanced intellectual skills are about to have their "aluminum moment," completely degrading from expensive luxury goods to cheap commodities. Aluminum, once more valuable than gold, became disposable due to the popularization of electrolysis, and AI's instant access to PhD-level knowledge will have the same effect. This portends a brutal future: the barriers of specialized knowledge accumulated over years could collapse instantly, and intelligence itself will no longer command a scarcity premium.

When intelligence becomes universally common, human relationships and emotional connections will become the only true anchors of value for human society. Photography once pushed art from realism towards Impressionism, which expressed the soul; similarly, AI's quest for optimal efficiency often presents "alien spaces" that transcend human intuition. The final conclusion is counterintuitive yet profound: in a future where machines handle all work, only trust and emotion between people are the ultimate hard currency that cannot be mass-produced by machines.

If you could remember only one thing from this dialogue, what would it be?

Soon, intelligence that used to be valuable will become as cheap as plastic bags. What will truly keep you competitive in the future is not a brain that can solve difficult problems, but the emotions to understand others and build trust.

Models and Cognition

Andrej Karpathy: From Vibe Coding to Agent Engineering (OpenAI Founding Member)

The speaker, Andrej Karpathy, is the most influential "educator-scientist" in the AI community. A founding member of OpenAI, he later served as Director of AI at Tesla, overseeing the autonomous driving vision system. In 2024, he left Tesla to found the AI education company Eureka Labs. His video series on YouTube, explaining neural networks step-by-step, serves as introductory material for countless AI engineers. He coined key concepts like "Software 2.0" and "Vibe Coding."

Even top experts feel "left behind" by the AI wave, as the evolution of technology has moved from auxiliary tools to autonomous systems. The speaker found in early 2026 that he no longer needed to modify code blocks generated by AI; he could simply trust the system to complete complex tasks. The significance of this judgment is that when AI can achieve self-correction and closed-loop delivery, the "baseline" for developers, previously built on experience, is violently raised, and personal learning speed can hardly keep up with the shifting technical foundation.

Modern computing is entering the Software 3.0 era, where an LLM is essentially a new type of computer using context as its leverage. Software 1.0 involved writing code, 2.0 training weights, and 3.0 programming within context (the memory space where a model processes information) via prompting. This means installing software no longer requires writing complex compatibility scripts; you just "feed" a description to an agent. Precise spelling of details is no longer a core competency.

Many existing application architectures are becoming "redundant" because AI can now process data directly at the raw data layer. The speaker found that his painstakingly developed menu generation app became meaningless because the model could now perform pixel-level rendering overlay directly on the photo. This advocates a profound change: AI should not just accelerate old business logic; we must realize that the disappearance of the middle layer means many traditional product forms have lost their physical basis for existence.

AI capabilities are "jagged"; it only exhibits superhuman intelligence in areas that can be verified. A model can refactor a hundred thousand lines of code but might fail at a simple common sense task like counting the 'r's in "strawberry." This is because models are primarily strengthened through RL (Reinforcement Learning, a training method that uses reward signals to guide model evolution) in verifiable domains like math and code. This reminds us: we must constantly observe within the loop, watching out for weaknesses that lie outside the model's training distribution.

We are not building "animals" with intrinsic motivation but "summoning ghosts" within the data distribution. The peak intelligence of a model depends on the distribution of training data (e.g., adding lots of chess game data makes it a chess master), not on it generating some biological curiosity. This counterintuitive judgment points out that AI doesn't have true "understanding"; it just intensely optimizes specific circuits in statistical simulation. Therefore, users must learn to identify and avoid false abilities unsupported by data.

Agentic engineering is about maintaining the quality red lines of professional software while leveraging stochastic AI. This new engineering method requires developers, while coordinating powerful but unstable agents, to ensure the system doesn't produce security vulnerabilities. It advocates a new 10x engineer paradigm: the core of competition is no longer the speed of writing code yourself but the ability to efficiently drive a large cluster of agents to deliver high-quality results, like a director.

When machines take over trivial API details, the real premium for humans will shift towards aesthetics and mastery of the "specification." Developers no longer need to memorize the specific interface parameters of PyTorch (a deep learning framework) because these details will be handled by memory-proficient AI "interns." This predicts a counterintuitive future: fundamental principles and design taste are more durable than tool details; humans should transition from "code bricklayers" to decision-makers who define "what constitutes good design."

"Thinking" can be outsourced, but "understanding" is the only throttle bottleneck for humans in the age of cheap intelligence. Although AI can help us process and recompile vast amounts of information, it cannot decide for us "why build this" and "is this valuable." This advocates an ultimate conclusion: humans remain the only commander of the system, because only human consciousness can give purpose to the intelligent processing power. This holistic understanding cannot be replaced by algorithms.

If you could remember only one thing from this dialogue, what would it be?

When machines can do all the work and even think through all the details for you, the only skill you absolutely can't afford to lose is figuring out exactly what you want, and being able to tell if the machine is doing it right.

Greg Brockman: Human Attention is the New Bottleneck (OpenAI Co-founder)

The speaker, Greg Brockman, is the co-founder and President of OpenAI. Former CTO of Stripe, he co-founded OpenAI with Sam Altman in 2015 and is the core architect of the company's technology and infrastructure. Within OpenAI, Altman handles external matters (fundraising, public image, policy), while Brockman handles internal matters (technology, computing power, products). His engineer-style habit of personally writing code and staying up late to monitor releases is well-known in Silicon Valley.

Intelligence has become a standardized, resellable commodity, leading to insatiable pathological growth in demand for computing power. OpenAI's business model essentially involves buying or leasing computing power, transforming it into intelligence via models, and reselling it at a premium. Since the demand for problem-solving is infinite, GPU (Graphics Processing Unit) supply projections for 2026 approach near zero. The importance of this judgment is that AI is no longer just a software service but has evolved into a resource-based commodity business, where the physical supply of computing power directly determines the upper limit of civilization's intelligence.

The Scaling law (the empirical rule that model capabilities increase with more compute) is an empirical truth on a cosmic scale, with no sign of a "wall" yet. Although the basic concepts of neural networks originated in the 1940s, as long as massive computing power is continuously invested, the capabilities of models increase correspondingly and deterministically. This advocates a key view: technological stagnation will not occur in the short term; as long as capital and electricity are continuously invested, we will obtain more powerful intelligence, providing a fundamental logical support for the aggressive investments of tech giants.

From a functional perspective, we have completed 80% of the journey towards AGI (Artificial General Intelligence), as models already possess the closed-loop capability to execute tasks independently. A systems engineer handed a complex optimization scheme to a model; the model not only wrote the code but autonomously ran a Profiler (performance analysis tool) and performed multiple rounds of optimization based on feedback until the task was fully completed. This advocates a counterintuitive view: AGI is not a future moment but an ongoing process. AI has evolved from a "code-writing assistant" into a "colleague who can solve problems."

Context (the background information a model holds for a specific task) is replacing model algorithms as the most critical competitive frontier. The new tool Chronicle can record everything a user does on their computer in real-time, giving AI "memory," thus saving humans the time of repeatedly explaining context to the machine. The importance of this judgment is that for entrepreneurs, one-time model training is no longer the only moat