AGI takeoff has started, says Altman.
Sam Altman, CEO of OpenAI, asserts that humanity has passed the “event horizon” and is close to building super intelligence, despite the current lack of widespread robots or daily AI interaction. He believes the most challenging scientific work for reaching the singularity is complete, with significant progress already made in AI’s ability to drive scientific discovery and increase productivity. Altman outlines a timeline predicting the arrival of agents capable of real cognitive work by 2025, systems with novel insights by 2026, and potentially robots in the real world by 2027, leading to a future of abundant intelligence and energy that could unlock unprecedented human progress.
Points clés
- Sam Altman, CEO of OpenAI, states that humanity is past the “event horizon” and close to building super intelligence.
- Altman believes the hardest scientific work to reach the singularity, leading to systems like GPT-4 and GPT-3, is already accomplished.
- The singularity is defined as a hypothetical moment when AI and other technologies become so advanced that humanity undergoes a dramatic and irreversible change.
- Altman highlights the potential for enormous gains in quality of life through AI driving faster scientific progress and increased productivity.
- Self-improving artificial intelligence is seen as the fastest route to unlocking scientific discovery.
- Timeline predictions include the arrival of agents capable of real cognitive work in 2025 and systems that can figure out novel insights in 2026.
- Altman suggests robots capable of doing tasks in the real world may arrive in 2027, though widespread proliferation is not anticipated soon.
- By 2030, one person’s ability to get much more done than in 2020 will be a striking change.
- In the 2030s, intelligence and energy are predicted to become wildly abundant.
- An average ChatGPT query uses about 0.3 watt-hours of energy and 0.000085 gallons of water.
- Altman acknowledges that entire classes of jobs will disappear but believes the world will become so much richer that new policy ideas, like Universal Basic Income (UBI), can be seriously considered.
- Altman suggests that “idea guys” who previously relied on others to build their concepts will have their day in the sun with the advent of AI agents.
- Intelligence “too cheap to meter” is considered within grasp, despite OpenAI’s current pricing structure.
À retenir
So, according to Sam Altman, we’re basically on the express train to a future where AI does everything, jobs are optional (thanks, UBI!), and “idea guys” finally get their moment in the sun. Apparently, the hardest part is over, and the rest is just smooth, exponential sailing. Just remember to solve that “alignment problem” first, because nobody wants a super-intelligent social media feed exploiting your fear and anger while you’re trying to enjoy your newfound abundance. And hey, if all else fails, you can always just plug into the Matrix – apparently, that’s on the table too!
Sources
Quiz sur la vidéo: 5 questions
