At its recently held 2026 AI Ascent summit, Sequoia Capital announced the arrival of AGI from the perspective of capital and services. In other words, agents that can independently break down tasks, work continuously, and complete them without human intervention in business and service settings have now reached maturity. Earlier AI systems could often work continuously for only a few minutes; by May 2026, AI could already work for several hours at a time.
Beyond the longer periods of work and greater processing capability involved in completing complex tasks with Claude Code, OpenAI’s Codex can even work continuously for eight hours to finish an extremely complex engineering project, with substantial quality in what it delivers.
With intelligence helping us, the time needed for large-scale data analysis, expert-level reasoning, and long-term, highly demanding team projects is shrinking sharply, while the barriers to entry are falling at an astonishing pace. At present, almost all code inside Anthropic and OpenAI is generated by AI. One-person companies, or OPCs, built around highly capable individuals are taking shape faster, and the boundaries of capability are expanding substantially. One metaphor at the summit was particularly memorable: in the 1860s, aluminum was more precious than gold, and Americans used it to cap the Washington Monument. Then electrolysis arrived, and within 30 years aluminum had become candy wrappers and drinks cans. Bhler said that aluminum is human intelligence, and electrolysis is AI. PhD-level expertise and industry know-how accumulated over more than a decade are experiencing the same fate. Once something scarce can be reproduced at scale, it ceases to be valuable.
We keep asking where progress is taking us and what it means. Where do possibilities lie? When traditional standards are being eroded faster by the tide of the times, when yesterday’s rules no longer apply today, when value shifts, the boundaries of capability dissolve, and experience no longer means what it once did—how can we reach a tomorrow worth reaching?
Perhaps the answer lies not in grand predictions, but in this group of young people. They may not have the highest GPAs, but they have already begun using AI to rewrite the boundaries of their own abilities.
On May 8, 2026, OpenAI announced its Future Stars program, recognizing 26 young people and teams who frequently used ChatGPT during university. All were around 20 years old, and all were highly capable individuals empowered by AI.
Ethan, 22, and Jamie, 26, worked with AI to start a company called Icarus Robotics, which researches space robotics. They have already raised $6.1 million and are preparing for tests on the International Space Station in early 2027.
Nayel, Arhan, and Rushil, all under 20, created Wi-Find, a project that uses Wi-Fi signals to detect people trapped behind walls. The machine-learning model they invented with AI identifies subtle changes in signals as they pass through rubble, determining whether someone is underneath.
Zeyneb, 19, worked with local communities to collect more than 500 minutes of language data and oral histories, and developed an efficient algorithm requiring little data. With help from ChatGPT and other tools, her project, The Revive Project, produced the first structured linguistic record of the endangered dialect Romeyka, including translated oral histories.
Others have trained a semantic search engine that can now search more than 100 million telescope images…
Achievements like these at around 20 years old would be hard to imagine without an enormous infusion of resources. Yet in this era they are happening before our eyes, and such examples are becoming increasingly common.
We cannot help asking: when the boundaries of individual capability dissolve, does that bring erosion or expansion? Does it weaken individual competitiveness, or comprehensively upgrade what a person can do?
We believe this is not settled by reflection alone, nor is it a binary opposition. The replacement of one specialized ability does not mean the disappearance of diverse capabilities. As standards shift, other qualities come into view. What ultimately determines the result is how we respond; the ability to act is the best answer. And when immediate action receives immediate feedback from AI, and the cost of trying is driven ever lower, we gain a new way of working for the first time: start doing, then work out the answers to those questions.
And when is the most competitive time to start? The answer is now!
More evidence suggests that AI capabilities are approaching a threshold in certain areas, yet there is no sign of them stopping. What we have accomplished, however, is nowhere near that threshold. On the side of demand, vast possibilities remain to be created. The real barriers of the future will lie not in technology itself, but in discovering more needs earlier and in each person’s way of meeting them. The first of these is precisely the ability to redefine problems. Our true limitation is that we do not dare to imagine or act.
Against this backdrop, capabilities connected with action rather than specialist skills become much more prominent. At Sequoia Capital’s closed-door AI summit, leaders across fields generally considered it very important to develop ‘distinctly human’ capabilities. Communication, interpersonal skills, emotional care, value judgments, and cooperation thus come to the fore. These uniquely human abilities will become more precious in the AI era. They shape the speed and quality of our actions, as well as the specific difficulties we encounter in developing them. It is often through interacting with real people that real opportunities and progress emerge.
Aluminum’s journey from the top of a monument to a drinks can was indeed a loss of value. Yet it was precisely because aluminum became so cheap that the aviation industry could develop and the modern life we see all around us could be built. The transformation of intelligence is similar. The point is not its own value, but everything we accomplish with it.
Returning to the title: what kind of tomorrow is worth reaching? The answer is clear—a tomorrow actively made by the person you are now, through your ideas and actions. We believe there are many such tomorrows. After all, beyond every tomorrow lies another entirely new tomorrow…
