1940
Era 1: 1940s-1950s

The Foundations

Birth of the "Machine Brain"

The journey began with mathematical abstractions. In 1943, Warren McCulloch and Walter Pitts modeled the first artificial neurons. By 1950, Alan Turing asked the pivotal question, "Can machines think?" and proposed the Turing Test. The era peaked in 1956 at the Dartmouth Workshop, where the term "Artificial Intelligence" was formally coined by John McCarthy.

Next: The First Winter 👇
Era 2: 1960s-1970s

The First AI Winter

Optimism Meets Reality

Early AI like ELIZA (the first chatbot) sparked wild imagination, but the technology couldn't keep up with the hype. By the mid-70s, the Lighthill Report in the UK and DARPA funding cuts in the US led to a collapse in research support. Scientists realized that "common sense" and "processing power" were far harder to solve than first anticipated.

The Expert Rise 👇
Era 3: 1980s

Expert Systems Boom

Corporate Intelligence

AI returned through "Expert Systems"—software that mimicked the decision-making of human specialists. Digital Equipment Corporation's XCON saved millions, proving AI's commercial value. This decade also saw the revival of neural networks through the backpropagation algorithm, though they remained niche compared to symbolic logic.

The Second Winter 👇
Era 4: 1990s

Second Winter & Deep Blue

The Quiet Before the Storm

As expert systems reached their limits, a second cooling period arrived. However, behind the scenes, Moore's Law was fueling a quiet revolution. In 1997, IBM's Deep Blue defeated World Chess Champion Garry Kasparov, a watershed moment that proved specialized machines could surpass human peak performance in structured environments.

The Data Explosion 👇
Era 5: 2000s

The ML Era

Big Data Rises

The internet changed everything. Suddenly, massive datasets were available to train machines. Instead of hard-coding rules, researchers like Geoffrey Hinton focused on "Deep Learning." In 2011, IBM's Watson won Jeopardy!, demonstrating that machines could handle the nuances of human language and trivia at scale.

The Neural Surge 👇
Era 6: 2010s

Deep Learning Explosion

The Neural Net Returns

2012's AlexNet proved that GPUs could train deep neural networks to recognize images better than any human-coded algorithm. This sparked a gold rush. Google's AlphaGo beat Lee Sedol in 2016, and in 2017, the "Transformer" architecture was invented, paving the way for the Large Language Models we use today.

The GenAI Present 👇
Era 7: 2020-2026

Generative AI & Agents

The Autonomous Future

We've moved beyond narrow search. Modern AI now plans, reasons, and executes complex tasks across all modalities. From LLMs that pass the Bar Exam to agents that can manage entire software projects autonomously, we are entering the age of AGI debates and universal AI assistance.

← Go Back

Foundation Models

The world's most powerful LLMs in 2026.

OpenAI

GPT-4o

Omni-modal reasoning at human speed. The new standard for AI interaction.

MMLU: 88.7% | click to expand
Anthropic

Claude 3.5 Opus

The subtle artist of the LLM world, excelling in code and creative reasoning.

MMLU: 89.1% | click to expand
Google

Gemini 2.0 Ultra

Unlimited windows into data. Native multimodality for the next generation.

MMLU: 90.0% | click to expand
Meta

Llama 3.1 405B

Bringing massive power to the open weights community. Privacy meets performance.

MMLU: 86.5% | click to expand
Creative Models 👇

Creative & Specialized

Image

Midjourney v7

The gold standard for AI art. Unmatched aesthetic understanding and detail.

Highest Aesthetic | click to expand
Video

Sora v2

Cinematic video generation with a deep understanding of physical physics.

State-of-the-Art | click to expand
Voice

ElevenLabs v3

Emotive, human-like speech synthesis and instant high-fidelity dubbing.

Natural Audio | click to expand
Science

AlphaFold 3

Solving the structure of all life's molecules with high precision.

Nobel Prize Tier | click to expand
Autonomous Agents 👇

Autonomous Agents

Cognition

Devin

Fully autonomous AI software engineer.

Browser

MultiOn

Navigate and act on any website automatically.

General

Auto-GPT v2

The foundation of iterative task fulfillment.

Task

BabyAGI

Dynamic goal prioritization and execution.

Hardware & Enterprise 👇

Hardware & Enterprise

Video

Sora v2

60-second cinematic clips with physical world logic.

Physically Informed | click to expand
Cognition

Devin

The first AI software engineer that can learn and build alone.

Autonomous SWE | click to expand
Browser

MultiOn

Navigate and act on any website automatically using LLM-powered browser automation.

Click to expand
General

Auto-GPT v2

The foundation of iterative task fulfillment, now with better memory and reasoning loops.

Click to expand