Topic
AGI
The leaders of the three frontier labs broadly agree that AGI is coming, but they disagree sharply on when and what it even means. Dario Amodei puts himself at 90% confidence that a "country of geniuses in a data center" arrives within ten years, with a personal hunch of one to two years. Sam Altman tells students in Delhi that his one-year-old son will never know a world where he was smarter than a computer. Shane Legg, who has held a 50/50 bet on AGI by 2028 since 2009, defines it more carefully than the others: his "minimal AGI" is a system that handles all typical human cognitive tasks reliably, without the bizarre failures current models still produce. Mira Murati, speaking at WIRED in late 2024 after leaving OpenAI, offered a wider range: achievable in a couple of years, or perhaps a decade or two. Ilya Sutskever skips the timeline debate entirely and argues from first principles... the brain is a biological computer, so a digital one can do everything it does. The when matters less to him than the what-then.
The real fracture is over what follows arrival. Dario warns the economic disruption will outrun society's ability to adapt, comparing the speed to previous transitions that took decades or centuries but will now happen in single-digit years. Jack Clark expects a political movement to freeze jobs in "bureaucratic amber" as a panicked response, and worries that protected busywork won't provide the meaning people need. Altman leans optimistic, framing super intelligence as an abundance engine that will make GDP a useless metric. Hassabis stays closest to concrete applications: he talks less about timelines and more about what AI will do for drug discovery and scientific research, treating AGI as a tool rather than an event. Sutskever calls it "the greatest challenge of humanity ever" and says most people, even those building it, struggle to internalize what's coming on an emotional level. Nobody in this group is a skeptic. The debate is between those who think we have a few years to prepare and those who think we have months.
People on this topic
Perspectives
Sutskever vs. the Scaling Consensus
Ilya Sutskever arguably proved scaling laws work better than anyone alive. He co-authored AlexNet, co-founded OpenAI, and oversaw the GPT series that turned neural scaling from a research curiosity into the dominant paradigm. Then he left and announced that the age of pure scaling is over. His argument: scaling sucked the air out of genuine research. Models trained on massive compute ace benchmarks but fail at generalization, like a student who memorized 10,000 competitive programming problems but can't architect real software. He thinks RL training produces meta-reward-hacking, where the researchers themselves (not just the models) are unconsciously overfitting to evaluation metrics. This puts him directly at odds with his former colleagues. Altman and Brockman are spending hundreds of billions on data centers and custom chips, betting that more compute is the binding constraint. Amodei still describes intelligence as a chemical reaction with known ingredients. Even Hassabis, who shares some of Sutskever's skepticism about pure scaling, is building gigawatt-scale infrastructure. Sutskever's counter-thesis is that AI needs something analogous to human emotions: a learned value function that provides fast, approximate feedback about whether a course of action is promising, rather than the sparse end-of-trajectory reward signals that current RL relies on. He cites a neurological case study of a stroke patient who lost emotional processing but retained full IQ, and became unable to make even trivial decisions. The implication: without an internal compass, raw intelligence is paralyzed. SSI is his bet that ideas, not compute budgets, are what's actually missing.
Hassabis vs. Everyone: What Counts as AGI
Demis Hassabis has the strictest definition of AGI among the frontier lab leaders, and it's not close. His test: train a model with a knowledge cutoff of 1911 and see if it can derive general relativity by 1915. Current systems cannot do this. He calls today's models "jagged intelligences" that win math olympiad gold but fail elementary arithmetic when the question is phrased differently. He identifies three specific capabilities still missing: continual learning after deployment, coherent long-term planning over years rather than minutes, and consistency across difficulty levels within the same domain. He also notes that general models play chess below weak amateur level, which he considers disqualifying. This stands in sharp contrast to the more expansive definitions others use. Amodei's "country of geniuses in a data center" is an economic metaphor, not a cognitive test. Altman's framing is generational (his son will never be smarter than a computer) rather than technical. Legg's three-tier spectrum (minimal, full, super) is more structured but sets the bar lower at the entry level: minimal AGI is just the point where failures stop being surprising. The practical consequence of Hassabis's stricter definition is that DeepMind invests differently. Rather than racing to declare AGI first, Hassabis bets on specialized tools (AlphaFold, AlphaProteo) orchestrated by general systems (Gemini), arguing that shoving all capabilities into one model degrades performance. His biggest play isn't a chatbot at all. It's Isomorphic Labs, which aims to compress drug discovery from a decade to weeks. That's a bet that AGI's value will be proven in scientific applications, not in passing cognitive benchmarks that the other labs are already designing their models to ace.