This site demonstrates one possible use of this domain. For acquisition, partnership, or investment inquiries, please use our contact link. (brainmatter.com)

timeline

AGI Development Timeline

Key milestones, theories, and capability jumps on the path toward artificial general intelligence.

A focused timeline of the conceptual and empirical milestones most relevant to AGI - the architectures, papers, and capability events that the field consistently cites.

1950Turing test

Alan Turing proposes the imitation game as an operational criterion for machine intelligence.

1956Dartmouth workshop

AI is founded as a research field with explicit AGI ambitions.

1980Chinese Room

Searle's thought experiment challenges the equivalence of symbol manipulation and understanding.

1997Deep Blue beats Kasparov

First AI to defeat a reigning world chess champion - narrow, but symbolically pivotal.

2005Kurzweil's Singularity

'The Singularity Is Near' popularizes the AGI/superintelligence forecast in mainstream discourse.

2012AlexNet

Deep learning becomes the dominant paradigm for perception.

2014Generative Adversarial Networks

Goodfellow et al. enable high-fidelity generative modeling.

2016AlphaGo defeats Lee Sedol

Deep RL + search masters a domain considered out of reach for decades.

2017Transformer

Self-attention architecture unlocks language-scale generalization.

2020GPT-3

175B-parameter LLM demonstrates broad few-shot capability - first 'frontier' model.

2022InstructGPT / RLHF

Alignment with human preferences makes LLMs usable as general assistants.

2022Emergent abilities documented

Wei et al. characterize capability jumps with scale.

2023GPT-4 / Claude 2

Multimodal frontier models reach expert-level on broad benchmarks.

2024o1 / reasoning models

Inference-time deliberation produces step-change gains on math, science, code.

2024ARC-AGI breakthrough

o3 reaches near-human performance on fluid reasoning benchmark.

2025Agentic systems

Long-horizon autonomous agents become production-deployable on real software tasks.

AGI Development Timeline - BRAINMATTER

Key takeaways

  • AGI progress is not linear - it advances in jumps coupled with architectural and training innovations.
  • Each new paradigm (search, deep learning, transformers, RLHF, reasoning) has redefined what 'intelligence' means in practice.
  • Capability and alignment research are now tightly coupled.

More visual knowledge