01:05:07
In an exclusive interview, OpenAI CEO Sam Altman discusses the revolutionary capabilities of GPT-5, the path to superintelligence, and how society might adapt to unprecedented technological change.
GPT-5 represents a significant leap beyond its predecessor with three core advancements:
Altman notes that while GPT-4 outperformed 90% of humans on standardized tests, GPT-5's real power lies in enabling rapid creative expression. "The speed with which you can bring ideas to life is remarkable," he observes, while acknowledging concerns about reduced "cognitive time under tension" in human learning.
When asked when AI might make significant scientific breakthroughs, Altman predicts:
"Most people will agree that significant discoveries will happen between 2025-2027. We've progressed from solving high school math problems to achieving IMO gold medal-level performance. The next leap requires moving from minute-long tasks to thousand-hour research cycles."
He envisions AI systems eventually designing experiments, interpreting results, and guiding drug discovery pipelines - potentially revolutionizing fields like cancer research through iterative human-AI collaboration.
Altman defines superintelligence practically:
"A kid born today will never be smarter than AI," Altman states, emphasizing this generation will consider intelligent machines as natural as smartphones.
Regarding viral AI-generated content like "bunnies on a trampoline," Altman suggests:
Addressing predictions that AI could displace half of entry-level white-collar jobs:
"Young people entering the workforce now are the luckiest in history. For the first time, one person can build a billion-dollar company using tools that replace what previously required hundreds."
He contrasts this with greater challenges facing older workers needing retraining, suggesting potential need for reimagined social contracts around resource distribution.
Timeframe | Capability | Impact |
---|---|---|
2025 (GPT-5) | Enhanced diagnostic accuracy | Reduced hallucinations in medical advice, improved symptom analysis |
2035 | AI-driven treatment discovery | Significant progress curing previously untreatable diseases |
Future | End-to-end research automation | AI designing experiments, analyzing results, and guiding clinical trials |
Altman identifies four critical development pillars:
Regarding responsibility for AI's societal impact, Altman advocates:
He draws parallels to transistor development - initially revolutionary, eventually ubiquitous and unnoticed in daily life.
Altman's core advice: "Get fluent with the tools. The most common question I get is how to prepare for AI, yet many haven't moved beyond using it as a better Google search."
He encourages hands-on experimentation to understand AI's capabilities, while acknowledging the need for personal resilience during rapid technological change.