The case for and against AGI by 2030 (article by Benjamin Todd)
May 12, 2025
auto_awesome
In this illuminating discussion, Benjamin Todd, a writer focused on AGI since 2014, breaks down the trends shaping the future of AI. He explores four key drivers of AI progress, including enhanced reasoning capabilities and the growing computational power fueling development. Todd contrasts the optimistic scenarios where AGI could emerge by 2030 and revolutionize industries like software and research with the challenges that might hinder such advancements. It's a thoughtful examination of the promising yet complex road ahead for artificial intelligence.
Four key drivers—scaling pre-training, reinforcement learning, increased reasoning time, and robust agent frameworks—are propelling AI advancements towards potential AGI by 2030.
The achievement of AGI, or similar advanced systems, could lead to transformative impacts on industries including software engineering and scientific research.
Despite the optimism surrounding AI capabilities, challenges such as regulatory hurdles and task complexity may hinder progress towards true AGI.
Deep dives
Growing Confidence in AGI Development
Recent statements from leading AI executives indicate a heightened level of confidence regarding the timeline of Artificial General Intelligence (AGI), with some suggesting that we could see its arrival by 2030. For example, OpenAI's Sam Altman has shifted from expressing uncertainty about the speed of development to claiming that they understand how to build AGI within a short timeframe. This change could result from four key drivers of progress in AI: larger base models, better reasoning methodologies through reinforcement learning, increased computational power, and enhanced agent scaffolding for complex tasks. Although some leaders in AI might be overestimating the timeline, the evidence supporting their confidence warrants serious consideration of the potential for rapid advancements in AI capabilities.
Key Drivers of AI Progress
Four critical drivers seem to underpin the current rapid advances in AI: scaling pre-training, reinforcement learning, extended reasoning time for models, and building robust agent frameworks. Scaling pre-training involves utilizing more computational resources to create base models that exhibit fundamental intelligence. Reinforcement learning teaches models to reason and solve complex problems by rewarding them for correct understandings, effectively enhancing their performance over time. These advancements are predicted to lead to AI systems that will significantly outperform humans in both coding and scientific reasoning by 2028, creating a foundational shift in how AI can assist with various tasks and industries.
The Potential of Reasoning Models
The introduction of reasoning models marks a pivotal moment in AI progression, allowing models to approach problem-solving more effectively. Reinforcement learning from human feedback has been crucial for teaching models to create logical chains of thought and tackle problems like scientific inquiries that require in-depth reasoning. In recent examples, models have demonstrated the ability to reach PhD-level performance in answering high-level scientific questions through these advanced reasoning techniques. This leap in capabilities underscores the likelihood that AI will not only excel in defined tasks but could also generate new insights, significantly impacting fields such as scientific research and technology development.
Implications of Extended Thinking Time
The enhancement of how long AI models can engage with problems is another significant aspect of their evolving capabilities. Models have shown improved accuracy by increasing their reasoning time, with projections suggesting that they will soon think for substantially longer durations, thereby yielding more detailed and accurate results. This ability will facilitate brute-force approaches to problem-solving and enable models to tackle complex tasks that require time and reflection. If this trend continues, we can expect AI to undertake increasingly elaborate tasks independently, moving us closer to achieving functionalities that align with AGI.
Transformative Economic Impact and Future Scenarios
As AI capabilities advance towards AGI, the economic ramifications could be profound, particularly in sectors reliant on knowledge work, such as software engineering and scientific research. The current trajectory suggests that AI could feasibly automate substantial portions of work, leading to unprecedented economic growth and innovation. However, challenges remain, including regulatory hurdles, the quality of training data, and the complexity of tasks that AI can manage. The next five years will be crucial in determining whether these advancements can sustain momentum, either producing transformative influences by 2030 or leading to a deceleration in progress due to emerging bottlenecks.
More and more people have been saying that we might have AGI (artificial general intelligence) before 2030. Is that really plausible?
This article by Benjamin Todd looks into the cases for and against, and summarises the key things you need to know to understand the debate. You can see all the images and many footnotes in the original article on the 80,000 Hours website.
In a nutshell:
Four key factors are driving AI progress: larger base models, teaching models to reason, increasing models’ thinking time, and building agent scaffolding for multi-step tasks. These are underpinned by increasing computational power to run and train AI systems, as well as increasing human capital going into algorithmic research.
All of these drivers are set to continue until 2028 and perhaps until 2032.
This means we should expect major further gains in AI performance. We don’t know how large they’ll be, but extrapolating recent trends on benchmarks suggests we’ll reach systems with beyond-human performance in coding and scientific reasoning, and that can autonomously complete multi-week projects.
Whether we call these systems ’AGI’ or not, they could be sufficient to enable AI research itself, robotics, the technology industry, and scientific research to accelerate — leading to transformative impacts.
Alternatively, AI might fail to overcome issues with ill-defined, high-context work over long time horizons and remain a tool (even if much improved compared to today).
Increasing AI performance requires exponential growth in investment and the research workforce. At current rates, we will likely start to reach bottlenecks around 2030. Simplifying a bit, that means we’ll likely either reach AGI by around 2030 or see progress slow significantly. Hybrid scenarios are also possible, but the next five years seem especially crucial.
Chapters:
Introduction (00:00:00)
The case for AGI by 2030 (00:00:33)
The article in a nutshell (00:04:04)
Section 1: What's driven recent AI progress? (00:05:46)
How we got here: the deep learning era (00:05:52)
Where are we now: the four key drivers (00:07:45)
Driver 1: Scaling pretraining (00:08:57)
Algorithmic efficiency (00:12:14)
How much further can pretraining scale? (00:14:22)
Driver 2: Training the models to reason (00:16:15)
How far can scaling reasoning continue? (00:22:06)
Driver 3: Increasing how long models think (00:25:01)
Driver 4: Building better agents (00:28:00)
How far can agent improvements continue? (00:33:40)
Section 2: How good will AI become by 2030? (00:35:59)
Trend extrapolation of AI capabilities (00:37:42)
What jobs would these systems help with? (00:39:59)
Software engineering (00:40:50)
Scientific research (00:42:13)
AI research (00:43:21)
What's the case against this? (00:44:30)
Additional resources on the sceptical view (00:49:18)
When do the 'experts' expect AGI? (00:49:50)
Section 3: Why the next 5 years are crucial (00:51:06)
Bottlenecks around 2030 (00:52:10)
Two potential futures for AI (00:56:02)
Conclusion (00:58:05)
Thanks for listening (00:59:27)
Audio engineering: Dominic Armstrong Music: Ben Cordell
Remember Everything You Learn from Podcasts
Save insights instantly, chat with episodes, and build lasting knowledge - all powered by AI.
