Rodney Brooks writes about the seven cognitive errors that lead to wildly mistaken predictions about the future of robotics and artificial intelligence. He frames four categories of predictions—Artificial General Intelligence, the Singularity, misaligned values, and evil AI entities—then dissects the reasoning failures behind them: overestimating short-term impact while underestimating long-term effects, treating far-future technology as unfalsifiable magic, confusing narrow task performance with broad competence, relying on ambiguous "suitcase words" like "learn" or "understand," extrapolating exponentials that will inevitably flatten, imagining Hollywood-style single-disruption scenarios, and ignoring the glacial pace of hardware deployment in the physical world.
- The OpenWorm project spent thirty years attempting to simulate C. elegans (302 neurons, 7,000 connections) bottom-up and was not yet halfway done as of 2017.
- iPod storage followed a perfect five-year exponential (10→160 GB) then collapsed abruptly once a single device could hold a complete music library.
- Modern factory automation still relies on PLCs introduced in 1968 that emulate electromechanical relays; Tesla was actively hiring PLC technicians at its Fremont factory.
- Autonomous vehicles drove on public roads in 1987 and coast-to-coast across the US in 1995, yet no large-scale deployment path had been identified by 2017.
- Brooks observes that humans coexisted with horses—autonomous agents with ongoing existences and super-human strength—for millennia without a single formal theorem about them.
- The OpenWorm project spent thirty years attempting to simulate C. elegans (302 neurons, 7,000 connections) bottom-up and was not yet halfway done as of 2017.
- iPod storage followed a perfect five-year exponential (10→160 GB) then collapsed abruptly once a single device could hold a complete music library.
- Modern factory automation still relies on PLCs introduced in 1968 that emulate electromechanical relays; Tesla was actively hiring PLC technicians at its Fremont factory.
- Autonomous vehicles drove on public roads in 1987 and coast-to-coast across the US in 1995, yet no large-scale deployment path had been identified by 2017.
- Brooks observes that humans coexisted with horses—autonomous agents with ongoing existences and super-human strength—for millennia without a single formal theorem about them.
AI safety and alignment research has predominantly been focused on methods for safeguarding individual AI systems, resting on the assumption of an eventual emergence of a monolithic Artificial General Intelligence (AGI). The alternative AGI emergence hypothesis, where general capability levels are first manifested through coordination in groups of sub-AGI individual agents with complementary skills and affordances, has received far less attention. Here we argue that this patchwork AGI hypothesis needs to be given serious consideration, and should inform the development of corresponding safeguards and mitigations.
Following Altman’s exit, Sutskever is tasked with guiding the team towards developing AI systems that not only push the boundaries of technology but also ensure they align with human values and safety protocols. His leadership style is expected to emphasize collaboration, transparency, and ongoing dialog with various stakeholders, including researchers, policymakers, and the public.
This article discusses Aigo's approach to achieving Artificial General Intelligence (AGI) through 'First Principles Analysis'. It argues that understanding the core principles of human intelligence – conceptual learning, adaptability, incremental learning, and metacognitive control – is crucial. It critiques current approaches like scaling LLMs and advocates for 'Cognitive AI' based on their Integrated Neuro-Symbolic Architecture (INSA).
DeepMind is prioritizing readiness, proactive risk assessment, and collaboration with the wider AI community as they explore the frontiers of AGI, focusing on mitigating risks like misuse and misalignment.
AAAI survey finds that most respondents are sceptical that the technology underpinning large-language models is sufficient for artificial general intelligence.
>"More than three-quarters of respondents said that enlarging current AI systems ― an approach that has been hugely successful in enhancing their performance over the past few years ― is unlikely to lead to what is known as artificial general intelligence (AGI). An even higher proportion said that neural networks, the fundamental technology behind generative AI, alone probably cannot match or surpass human intelligence. And the very pursuit of these capabilities also provokes scepticism: less than one-quarter of respondents said that achieving AGI should be the core mission of the AI research community.
>"More than three-quarters of respondents said that enlarging current AI systems ― an approach that has been hugely successful in enhancing their performance over the past few years ― is unlikely to lead to what is known as artificial general intelligence (AGI). An even higher proportion said that neural networks, the fundamental technology behind generative AI, alone probably cannot match or surpass human intelligence. And the very pursuit of these capabilities also provokes scepticism: less than one-quarter of respondents said that achieving AGI should be the core mission of the AI research community.
Google co-founder Sergey Brin has called for the company to 'turbocharge' its efforts in the race to achieve Artificial General Intelligence (AGI), citing increased competition and the need for more efficient use of AI in coding and productivity.
> "The other ingredient is a call for employees to double down on their work. This includes a recommendation of “being in the office at least every weekday” and that “60 hours a week is the sweet spot of productivity,” while warning that more might result in burnout. "
> "The other ingredient is a call for employees to double down on their work. This includes a recommendation of “being in the office at least every weekday” and that “60 hours a week is the sweet spot of productivity,” while warning that more might result in burnout. "
Zvi discusses AI, alignment, geopolitics, and more on Dwarksh Patel's podcast with Leopold Aschenbrenner. Topics include the timeline for AI development, the intelligence explosion, the geopolitics of AGI, and more.
