AI & ML Timeline 1943 – 2026
307 events · updated Oct 20, 2026
1943

A Logical Calculus of the Ideas Immanent in Nervous Activity — formalizing Cajal's neuron doctrine as computational units

Warren McCulloch; Walter Pitts (University of Chicago); Santiago Ramón y Cajal (foundational)

Research Breakthroughs
Details →
1945

As We May Think essay published

Vannevar Bush (OSRD)

Research Breakthroughs
Details →
1949

Hebbian learning introduced (*The Organization of Behavior*)

Donald Hebb (McGill)

Research Breakthroughs
Details →
1950

Turing Test described

Alan Turing (University of Manchester)

Research Breakthroughs
Details →
1951

SNARC early neural network hardware

Marvin Minsky; Dean Edmonds (Harvard)

Hardware & Infrastructure
Details →
1952

Checkers program (early AI game-playing)

Arthur Samuel (IBM)

Research Breakthroughs
Details →
1956

Term "artificial intelligence" coined; Logic Theorist program

John McCarthy (Dartmouth); Allen Newell; Herbert Simon; Cliff Shaw (RAND/CMU)

Research Breakthroughs
Details →
1957

"You shall know a word by the company it keeps" — distributional semantics (conceptual origin of embeddings)

J.R. Firth (linguistics)

Research Breakthroughs
Details →
1958

LISP programming language created

John McCarthy (MIT)

Enterprise & Developer Tools
Details →

Perceptron algorithm developed

Frank Rosenblatt (Cornell)

Research Breakthroughs
Details →
1959

"Machine learning" framing; self-learning checkers popularized

Arthur Samuel (IBM)

Research Breakthroughs
Details →
1960

ADALINE neural network

Bernard Widrow; Ted Hoff (Stanford)

Research Breakthroughs
Details →
1965

DENDRAL expert system project begins

Edward Feigenbaum; Joshua Lederberg; Bruce Buchanan (Stanford)

Research Breakthroughs
Details →

Fuzzy sets paper published

Lotfi Zadeh (UC Berkeley)

Research Breakthroughs
Details →
1966

Shakey robot project begins (mobile reasoning robot)

Charles Rosen; Nils Nilsson; Peter Hart (SRI, DARPA)

Autonomous Systems
Details →

ELIZA chatbot described in CACM

Joseph Weizenbaum (MIT)

Research Breakthroughs
Details →

ALPAC report slows machine translation funding

US National Academy of Sciences (ALPAC)

Governance & Society
Details →
1968

Visual cortex receptive-field discoveries (foundations for convnets)

David Hubel; Torsten Wiesel (Harvard)

Research Breakthroughs
Details →

A* search algorithm introduced

Peter Hart; Nils Nilsson; Bertram Raphael (SRI)

Research Breakthroughs
Details →
1969

Perceptrons book highlights limits of single-layer perceptrons (e.g., XOR)

Marvin Minsky; Seymour Papert (MIT)

Research Breakthroughs
Details →

Kaissa chess program developed

Soviet researchers (USSR)

Autonomous Systems
Details →
1970

SHRDLU natural-language blocks world

Terry Winograd (MIT)

Research Breakthroughs
Details →

Early continuous speech recognition work

Raj Reddy (CMU)

Research Breakthroughs
Details →
1971

ARPA issues RFP for Speech Understanding Research (SUR) program

Allen Newell (study group lead); ARPA

Governance & Society
Details →
1972

Prolog programming language introduced

Alain Colmerauer; Philippe Roussel (Aix-Marseille)

Enterprise & Developer Tools
Details →
1973

"Artificial Intelligence: A General Survey" (Lighthill report)

James Lighthill (UK Science Research Council)

Governance & Society
Details →
1974

First AI winter begins (funding and expectations crash)

ARPA; UK funders

Governance & Society
Details →
1975

Backpropagation dissertation

Paul Werbos (Harvard)

Research Breakthroughs
Details →

Frames introduced for knowledge representation

Marvin Minsky (MIT)

Research Breakthroughs
Details →
1976

MYCIN expert system

Bruce Buchanan; Edward Shortliffe (Stanford)

Research Breakthroughs
Details →
1977

Expectation-Maximization (EM) algorithm introduced

Dempster; Laird; Rubin

Research Breakthroughs
Details →
1979

Industrial robot fatality (safety & labor milestone)

Robert Williams (Ford)

Governance & Society
Details →
1980

Neocognitron (early convnet architecture)

Kunihiko Fukushima (NHK)

Research Breakthroughs
Details →

XCON expert system deployed at DEC

John McDermott (Carnegie Mellon); Digital Equipment Corporation (DEC)

AI Products
Details →
1982

Hopfield Network

John Hopfield (Caltech)

Research Breakthroughs
Details →

Dragon Systems founded (speech recognition company)

James Baker; Janet Baker (Dragon Systems)

AI Products
Details →
1983

Ironies of Automation published (foundational work on human-automation interaction)

Lisanne Bainbridge (University College London)

Governance & Society
Details →
1984

CART (Classification and Regression Trees)

Breiman; Friedman; Olshen; Stone

Research Breakthroughs
Details →
1986

Backpropagation popularized in modern neural nets

David Rumelhart; Geoffrey Hinton; Ronald Williams

Research Breakthroughs
Details →

Attention, Intentions, and the Structure of Discourse

Barbara Grosz; Candace Sidner

Research Breakthroughs
Details →

ID3 decision tree learning algorithm

Ross Quinlan (University of Sydney)

Research Breakthroughs
Details →

First robot car demonstrations

Ernst Dickmanns (Universität der Bundeswehr München)

Autonomous Systems
Details →

Boltzmann Machine developed

Terry Sejnowski; Geoffrey Hinton

Research Breakthroughs
Details →
1987

NETtalk learns to pronounce English text

Terry Sejnowski; Charles Rosenberg (Salk)

Research Breakthroughs
Details →
1988

Jabberwacky/Thoughts learning-chatbot project begins

Rollo Carpenter (UK)

AI Products
Details →

Probabilistic reasoning / Bayesian networks consolidated (textbook)

Judea Pearl (UCLA)

Research Breakthroughs
Details →

"A Statistical Approach to Language Translation"

IBM

Research Breakthroughs
Details →
1989

Time-delay neural network (TDNN) for speech

Alex Waibel (CMU)

Research Breakthroughs
Details →

Q-learning introduced

Chris Watkins (King's College Cambridge)

Research Breakthroughs
Details →

Convolutional neural networks for vision tasks

Yann LeCun (AT&T Bell Labs)

Research Breakthroughs
Details →
1990

Latent Semantic Analysis (SVD) maps word relationships — precursor to latent embedding space

Scott Deerwester; Susan Dumais; George Furnas; Thomas Landauer; Richard Harshman (Bellcore; University of Tennessee)

Research Breakthroughs
Details →

Neuromorphic electronic systems described

Carver Mead (Caltech)

Hardware & Infrastructure
Details →

"The Age of Intelligent Machines" published

Ray Kurzweil

Research Breakthroughs
Details →

Elman Networks (Simple Recurrent Networks) — first recurrent architecture with true context units, giving neural networks explicit sequence memory

Jeffrey Elman (UC San Diego)

Research Breakthroughs
Details →
1991

DART supply chain optimization used in Gulf War

DARPA

AI Products
Details →

"Intelligence without representation" (behavior-based robotics)

Rodney Brooks (MIT)

Autonomous Systems
Details →
1992

REINFORCE policy-gradient algorithm

Ronald Williams (Northeastern)

Research Breakthroughs
Details →
1993

Penn Treebank released — annotated corpus of Wall Street Journal text that became the gold standard for NLP parsing and tagging benchmarks for a decade

Mitchell Marcus; Beatrice Santorini; Mary Ann Marcinkiewicz (University of Pennsylvania)

Research Breakthroughs
Details →

Polly vision-enabled behavior-based robot

Ian Horswill (MIT)

Autonomous Systems
Details →

MIT Cog humanoid robotics project

Rodney Brooks; Lynn Stein; Cynthia Breazeal (MIT)

Autonomous Systems
Details →
1994

Byte Pair Encoding (BPE) introduced as data compression — later adopted for NLP tokenization

Philip Gage (compression)

Research Breakthroughs
Details →

"Soft computing" term popularized

Lotfi Zadeh (UC Berkeley)

Research Breakthroughs
Details →
1995

Random decision forests developed

Tin Kam Ho (Bell Labs)

Research Breakthroughs
Details →

Support Vector Machines introduced

Corinna Cortes; Vladimir Vapnik (AT&T Bell Labs)

Research Breakthroughs
Details →

Robot car drives 1000 miles Munich to Copenhagen

Ernst Dickmanns (Universität der Bundeswehr München)

Autonomous Systems
Details →
1996

AdaBoost introduced

Yoav Freund; Robert Schapire

Research Breakthroughs
Details →

TD-Gammon shows power of temporal-difference learning

Gerald Tesauro (IBM)

Research Breakthroughs
Details →
1997

Deep Blue defeats Garry Kasparov in chess

IBM

Autonomous Systems
Details →

Long Short-Term Memory (LSTM)

Sepp Hochreiter; Jürgen Schmidhuber

Research Breakthroughs
Details →
1998

Planning and acting in partially observable stochastic domains (POMDPs)

Leslie Kaelbling; Michael Littman; Anthony Cassandra (MIT/Brown)

Research Breakthroughs
Details →

LeNet-5 / CNNs for document recognition

Yann LeCun; Léon Bottou; Yoshua Bengio; Patrick Haffner (AT&T Bell Labs)

Research Breakthroughs
Details →

Reinforcement Learning textbook (widely influential)

Richard Sutton; Andrew Barto

Research Breakthroughs
Details →

Minerva and Pearl robots

Sebastian Thrun (CMU)

Autonomous Systems
Details →
1999

Sony AIBO consumer robot

Sony

Autonomous Systems
Details →
2000

KISMET emotional robot

Cynthia Breazeal (MIT)

Autonomous Systems
Details →

Neural Probabilistic Language Model — learns distributed word representations to combat curse of dimensionality (presented at NIPS 2000; published in JMLR 2003)

Yoshua Bengio; Réjean Ducharme; Pascal Vincent (Université de Montréal)

Foundation Models
Details →
2001

Echo state network developed

Herbert Jaeger (Jacobs_University_Bremen)

Research Breakthroughs
Details →

Random Forests introduced

Leo Breiman (UC Berkeley)

Research Breakthroughs
Details →

Conditional Random Fields (CRFs) introduced

John Lafferty; Andrew McCallum; Fernando Pereira

Research Breakthroughs
Details →
2002

iRobot Roomba autonomous floor cleaning robot

iRobot

Autonomous Systems
Details →
2003

Latent Dirichlet Allocation (LDA) topic modeling

David Blei; Andrew Ng; Michael Jordan

Research Breakthroughs
Details →
2004

DARPA Grand Challenge for autonomous vehicles

DARPA

Autonomous Systems
Details →
2005

Blue Brain Project launched (brain simulation neuroscience)

Henry Markram (EPFL)

Research Breakthroughs
Details →
2006

Deep Belief Nets (DBNs) revive deep learning

Geoffrey Hinton; Simon Osindero; Yee-Whye Teh (University of Toronto)

Research Breakthroughs
Details →
2007

CUDA introduced for general-purpose GPU computing

NVIDIA

Hardware & Infrastructure
Details →

GPUs enable large-scale neural network training (early deep learning acceleration)

NVIDIA; researchers

Hardware & Infrastructure
Details →

Stacked Auto-Encoders

Yoshua Bengio (Université_de_Montréal)

Research Breakthroughs
Details →

Checkers game solved (perfect play leads to draw)

Jonathan Schaeffer et al. (University of Alberta)

Autonomous Systems
Details →
2008

Neuromorphic computing program (SyNAPSE) launched

Dharmendra Modha (IBM)

Hardware & Infrastructure
Details →

Open Advancement of Question Answering (OAQA) initiative

IBM; Carnegie Mellon University

Enterprise & Developer Tools
Details →
2009

Deep Boltzmann Machines (DBMs)

Ruslan Salakhutdinov; Geoffrey Hinton

Research Breakthroughs
Details →

Robot Operating System (ROS) presented and broadly released

Morgan Quigley; Willow Garage; Stanford (STAIR)

Autonomous Systems
Details →

ImageNet dataset introduced

Jia Deng; Wei Dong; Richard Socher; Li-Jia Li; Kai Li; Fei-Fei Li (Princeton)

Research Breakthroughs
Details →

Google self-driving car project begins (internal)

Google; Sebastian Thrun (Stanford/Google)

Autonomous Systems
Details →
2010

DeepMind incorporated

Demis Hassabis; Shane Legg; Mustafa Suleyman (DeepMind)

Governance & Society
Details →

Rectified Linear Units (ReLU) activation function

Vinod Nair; Geoffrey Hinton (University of Toronto)

Research Breakthroughs
Details →
2011

IBM Watson defeats humans on Jeopardy!

IBM (DeepQA)

Research Breakthroughs
Details →

Siri digital assistant introduced

Apple

AI Products
Details →
2012

Google Now digital assistant launched

Google

AI Products
Details →

Deep neural networks transform speech recognition

Geoffrey Hinton; George Dahl; Abdel-rahman Mohamed

Research Breakthroughs
Details →

AlexNet wins ImageNet (deep CNN breakthrough)

Alex Krizhevsky; Ilya Sutskever; Geoffrey Hinton (University of Toronto)

Research Breakthroughs
Details →

Dropout regularization technique introduced

Geoffrey Hinton; Nitish Srivastava; Alex Krizhevsky; Ilya Sutskever; Ruslan Salakhutdinov (University of Toronto)

Research Breakthroughs
Details →
2013

Word2Vec: neural word embeddings (hidden layer as embedding)

Tomas Mikolov; Kai Chen; Greg Corrado; Jeffrey Dean (Google)

Research Breakthroughs
Details →

Variational Autoencoders (VAE) — introduced the learned latent space concept, enabling generative models to sample and interpolate in a continuous representation

Diederik Kingma; Max Welling (University of Amsterdam)

Generative AI
Details →
2014

DeepFace achieves near-human face verification

Yaniv Taigman; Ming Yang; Marc'Aurelio Ranzato; Lior Wolf (Facebook/TAU)

Research Breakthroughs
Details →

Adam optimizer introduced

Diederik Kingma; Jimmy Ba

Research Breakthroughs
Details →

Generative Adversarial Networks (GANs) introduced

Ian Goodfellow; Jean Pouget-Abadie; Mehdi Mirza; et al.

Generative AI
Details →

VGGNet demonstrates importance of network depth

Karen Simonyan; Andrew Zisserman (University of Oxford)

Research Breakthroughs
Details →

Seq2Seq learning with neural networks

Ilya Sutskever; Oriol Vinyals; Quoc Le (Google)

Research Breakthroughs
Details →

Gated Recurrent Units (GRU) — a streamlined alternative to LSTM that powered early neural machine translation and seq2seq models

Kyunghyun Cho; Bart van Merrienboer; Yoshua Bengio et al.

Research Breakthroughs
Details →

Neural machine translation attention mechanism

Dzmitry Bahdanau; Kyunghyun Cho; Yoshua Bengio

Research Breakthroughs
Details →

Microsoft Cortana digital assistant announced

Microsoft

AI Products
Details →

Amazon Echo/Alexa announced

Amazon (Lab126)

AI Products
Details →

Google acquires DeepMind

Google/DeepMind

Governance & Society
Details →
2015

OpenAI founded (nonprofit)

OpenAI

Governance & Society
Details →

DQN in Nature (human-level Atari control)

Volodymyr Mnih; Koray Kavukcuoglu; David Silver; et al. (DeepMind)

Research Breakthroughs
Details →

Batch Normalization accelerates deep network training

Sergey Ioffe; Christian Szegedy (Google)

Research Breakthroughs
Details →

ResNet enables very deep CNNs

Kaiming He; Xiangyu Zhang; Shaoqing Ren; Jian Sun (Microsoft Research)

Research Breakthroughs
Details →

TensorFlow open-sourced

Google

Enterprise & Developer Tools
Details →

AlphaGo defeats a pro player (Fan Hui match; announced 2016)

DeepMind

Autonomous Systems
Details →

Google self-driving cars surpass 1M autonomous miles

Google (Self-Driving Car Project)

Autonomous Systems
Details →

Deep Unsupervised Learning using Nonequilibrium Thermodynamics — introduced the forward/reverse diffusion process as a generative framework, the conceptual origin of modern diffusion models

Jascha Sohl-Dickstein; Eric Weiss; Niru Maheswaranathan; Surya Ganguli (Stanford; Google Brain)

Research Breakthroughs
Details →
2016

BPE adopted for NLP tokenization (subword units for rare words in neural machine translation)

Rico Sennrich; Barry Haddow; Alexandra Birch (University of Edinburgh)

Research Breakthroughs
Details →

AlphaGo defeats Lee Sedol (world champion)

DeepMind

Autonomous Systems
Details →

YOLO (You Only Look Once) real-time object detection

Joseph Redmon et al.

Research Breakthroughs
Details →

WaveNet generative model for raw audio

DeepMind

Generative AI
Details →
2017

PyTorch open-sourced

Meta (Facebook AI Research)

Enterprise & Developer Tools
Details →

Graph Convolutional Networks introduced

Thomas Kipf; Max Welling

Research Breakthroughs
Details →

Federated Learning enables privacy-preserving distributed ML

H. Brendan McMahan et al. (Google)

Research Breakthroughs
Details →

Libratus defeats professional poker players

Tuomas Sandholm; Noam Brown (CMU)

Autonomous Systems
Details →

"Attention is all you need" paper introduces Transformers

Ashish Vaswani et al. (Google Brain)

Foundation Models
Details →

Google TPU (Tensor Processing Unit) announced

Google

Hardware & Infrastructure
Details →

NVIDIA Volta V100 — first GPU with Tensor Cores (dedicated hardware matrix-multiply units), making Transformer-scale training practically feasible

NVIDIA

Hardware & Infrastructure
Details →

Capsule Networks introduced

Geoffrey Hinton

Research Breakthroughs
Details →
2018

Alibaba language AI beats humans in reading test

Alibaba

Foundation Models
Details →

ELMo introduces deep contextualized word representations

Matthew Peters et al. (Allen Institute for AI)

Foundation Models
Details →

Google Duplex conversational AI demonstrated

Google

Foundation Models
Details →

GPT (Generative Pre-trained Transformer) released

Alec Radford et al. (OpenAI)

Foundation Models
Details →

GLUE benchmark — unified multi-task NLP evaluation suite that defined the BERT era and drove rapid progress in language understanding

Alex Wang et al. (NYU; University of Washington; DeepMind)

Research Breakthroughs
Details →

BERT (Bidirectional Encoder Representations from Transformers)

Jacob Devlin et al. (Google AI Language)

Foundation Models
Details →
2019

GPT-2 released (1.5B parameters); concerns about misuse

Alec Radford et al. (OpenAI)

Foundation Models
Details →

DeepMind's AlphaStar reaches Grandmaster level in StarCraft II

DeepMind

Autonomous Systems
Details →

PyTorch 1.0 production release

Meta (Facebook AI Research)

Enterprise & Developer Tools
Details →

Nvidia A100 GPU released

Nvidia

Hardware & Infrastructure
Details →

StyleGAN — photorealistic face synthesis via style-based generator; "This Person Does Not Exist" demonstrated GANs could reach near-perfect perceptual realism

Tero Karras; Samuli Laine; Timo Aila (NVIDIA)

Generative AI
Details →

Score-based generative modeling via estimated gradients of the data distribution using annealed Langevin dynamics — foundational work connecting score functions to generative modeling

Yang Song; Stefano Ermon (Stanford University)

Research Breakthroughs
Details →
2020

GPT-3 released (175B parameters)

Tom Brown et al. (OpenAI)

Foundation Models
Details →

AlphaFold 2 solves protein folding problem

John Jumper et al. (DeepMind)

Research Breakthroughs
Details →

Denoising Diffusion Probabilistic Models (DDPM) — connected diffusion to denoising score matching, enabling high-quality image synthesis and making diffusion models practical

Jonathan Ho; Ajay Jain; Pieter Abbeel (UC Berkeley)

Research Breakthroughs
Details →

DALL-E 1 announced

Aditya Ramesh et al. (OpenAI)

Generative AI
Details →

The Pile — 800GB open-source training corpus enabling non-corporate LLM research

Leo Gao et al. (EleutherAI)

Research Breakthroughs
Details →

Switch Transformer (1.6T parameters)

William Fedus et al. (Google)

Foundation Models
Details →
2021

GitHub Copilot launched publicly

GitHub/OpenAI

AI Products
Details →

AlphaFold 2 paper published in Nature

DeepMind

Research Breakthroughs
Details →

Codex released (powers GitHub Copilot)

OpenAI

Enterprise & Developer Tools
Details →

Wu Dao 2.0 (1.75T parameters)

Beijing Academy of AI

Foundation Models
Details →

Tesla AI Day showcases Dojo supercomputer

Tesla

Hardware & Infrastructure
Details →

Meta announces AI Research SuperCluster (RSC)

Meta

Hardware & Infrastructure
Details →

Structured State Spaces (S4) — first sequence model to match Transformers on long-range tasks; direct precursor to Mamba

Albert Gu; Karan Goel; Christopher Ré (Stanford University)

Research Breakthroughs
Details →
2022

ChatGPT launched (November 30) - fastest growing consumer app

OpenAI

Foundation Models
Details →

High-Resolution Image Synthesis with Latent Diffusion Models — diffusion in learned latent space with cross-attention conditioning; technical foundation for Stable Diffusion

Robin Rombach; Andreas Blattmann; Dominik Lorenz; Patrick Esser; Björn Ommer (University of Heidelberg)

Research Breakthroughs
Details →

DALL-E 2 paper — hierarchical text-conditional image generation using diffusion in CLIP embedding space, demonstrating strong text-image alignment

Aditya Ramesh; Prafulla Dhariwal; Alex Nichol; Casey Chu; Mark Chen (OpenAI)

Generative AI
Details →

Imagen — photorealistic text-to-image diffusion using large pre-trained T5 language models for text encoding; introduced DrawBench evaluation benchmark

Chitwan Saharia et al. (Google Research)

Generative AI
Details →

DALL-E 2 released publicly

OpenAI

Generative AI
Details →

Stable Diffusion released (open source)

Stability AI

Generative AI
Details →

Midjourney v3 and v4 released

Midjourney

Generative AI
Details →

Galactica (science-focused LLM) released and quickly withdrawn

Meta

Foundation Models
Details →

LaMDA controversy (sentience claims)

Google/Blake Lemoine

Governance & Society
Details →

Whisper (speech recognition) released

OpenAI

Foundation Models
Details →

AlphaTensor discovers faster matrix multiplication

DeepMind

Research Breakthroughs
Details →

Joint Embedding Predictive Architecture (JEPA) introduced

Yann LeCun (Meta)

Research Breakthroughs
Details →

PaLM (540B parameters)

Google

Foundation Models
Details →
2023

GPT-4 released (multimodal) in March

OpenAI

Foundation Models
Details →

ChatGPT reaches 100M users (fastest ever)

OpenAI

AI Products
Details →

Google releases Bard (later renamed Gemini)

Google

Foundation Models
Details →

Claude released by Anthropic

Anthropic (Dario Amodei)

Foundation Models
Details →

LLaMA and LLaMA 2 released (open source)

Meta

Foundation Models
Details →

Mistral AI founded; releases Mistral 7B

Mistral AI

Foundation Models
Details →

Microsoft Copilot integrated across Office 365

Microsoft

AI Products
Details →

Adobe Firefly (generative AI for creative tools)

Adobe

Generative AI
Details →

Runway Gen-2 (text-to-video)

Runway

Generative AI
Details →

AutoGPT and BabyAGI (autonomous agents)

Open source community

Autonomous Systems
Details →

Code Interpreter (Advanced Data Analysis) for ChatGPT

OpenAI

AI Products
Details →

Sam Altman fired and rehired from OpenAI (November)

OpenAI/Microsoft

Governance & Society
Details →

EU AI Act passed

European Union

Governance & Society
Details →

GPT-4 Vision (GPT-4V) released

OpenAI

Generative AI
Details →

Google Gemini announced (December)

Google

Foundation Models
Details →

Mamba: Linear-Time Sequence Modeling with Selective State Spaces

Albert Gu and Tri Dao

Research Breakthroughs
Details →

Nvidia H100 becomes standard for AI training

Nvidia

Hardware & Infrastructure
Details →

Character.AI reaches millions of users

Character.AI

AI Products
Details →

ChatGPT gets custom instructions and GPTs

OpenAI

AI Products
Details →

Meta AI launched with celebrity avatars

Meta

AI Products
Details →

X.AI launches Grok

Elon Musk/xAI

Foundation Models
Details →
2024

Sora (text-to-video) announced

OpenAI

Generative AI
Details →

Claude 3 family released (Opus, Sonnet, Haiku)

Anthropic

Foundation Models
Details →

Gemini 1.5 with 1M token context window

Google

Foundation Models
Details →

GPT-4o (omni) released - faster multimodal

OpenAI

Generative AI
Details →

Google I/O: Gemini integration across products

Google

AI Products
Details →

Apple Intelligence announced (June WWDC)

Apple

AI Products
Details →

ChatGPT desktop app and GPT-4o mini released

OpenAI

AI Products
Details →

Anthropic Claude raises $4B from Amazon

Anthropic/Amazon

Governance & Society
Details →

Llama 3 and Llama 3.1 (405B parameters) released

Meta

Foundation Models
Details →

Mistral Large 2 released

Mistral AI

Foundation Models
Details →

DeepSeek-V2 (Chinese model competitive with GPT-4)

DeepSeek

Foundation Models
Details →

Perplexity AI gains traction as AI search engine

Perplexity

AI Products
Details →

NotebookLM with Audio Overview feature

Google

AI Products
Details →

ChatGPT Search launched

OpenAI

AI Products
Details →

California SB 1047 (AI safety bill) vetoed

California Legislature

Governance & Society
Details →

Strawberry (OpenAI o1) - reasoning model released

OpenAI

Foundation Models
Details →

Grok 2 with image generation

xAI

Foundation Models
Details →

Amazon Nova models announced

Amazon

Foundation Models
Details →

Google Gemini 2.0 Flash released

Google

Foundation Models
Details →

OpenAI announces o1 and o1-pro (December)

OpenAI

Foundation Models
Details →

Microsoft Copilot Vision announced

Microsoft

Generative AI
Details →

Runway Gen-3 Alpha released

Runway

Generative AI
Details →

ElevenLabs reaches unicorn status

ElevenLabs

Generative AI
Details →

Anthropic publishes research on AI interpretability

Anthropic

Governance & Society
Details →

Figure 01 humanoid robot with ChatGPT integration

Figure AI/OpenAI

Autonomous Systems
Details →

Tesla Optimus Gen 2 demonstrated

Tesla

Autonomous Systems
Details →

Nvidia becomes most valuable company (briefly)

Nvidia

Hardware & Infrastructure
Details →

Claude 3.5 Sonnet released

Anthropic

Foundation Models
Details →

GPT-4o with advanced voice mode

OpenAI

Generative AI
Details →

OpenAI DevDay announcements (realtime API)

OpenAI

Enterprise & Developer Tools
Details →

Anthropic Computer Use feature (Claude controls computers)

Anthropic

Autonomous Systems
Details →

ChatGPT reaches 200M weekly active users

OpenAI

AI Products
Details →

Waymo expands autonomous taxi service significantly

Waymo/Alphabet

Autonomous Systems
Details →

AlphaFold 3 released

Google DeepMind

Research Breakthroughs
Details →

Sora video generation tool released publicly

OpenAI

Generative AI
Details →

DeepSeek-V3 released (671B MoE, trained for ~$5.5M)

DeepSeek

Foundation Models
Details →

ChatGPT Pro subscription ($200/month) launched

OpenAI

AI Products
Details →
2025

DeepSeek R1 reasoning model (open source)

DeepSeek

Foundation Models
Details →

Claude Opus 4 and Sonnet 4 released

Anthropic

Foundation Models
Details →

Gemini 2.0 widely available

Google

Foundation Models
Details →

Deep research features in ChatGPT and other platforms

Multiple

AI Products
Details →

Operator (browser agent) announced by OpenAI

OpenAI

Autonomous Systems
Details →

Gemini Robotics brings AI into physical world

Google DeepMind

Autonomous Systems
Details →

WeatherNext 2 advanced weather forecasting

Google

Research Breakthroughs
Details →

SIMA 2 AI agent for 3D worlds

Google DeepMind

Autonomous Systems
Details →

Genie 3 interactive world model

Google DeepMind

Generative AI
Details →

NVIDIA Blackwell architecture GPUs launch

NVIDIA

Hardware & Infrastructure
Details →

EU AI Act first prohibitions enforced — unacceptable-risk AI systems banned

European Union

Governance & Society
Details →

Grok 3 released

xAI

Foundation Models
Details →

Claude Code agentic coding CLI released

Anthropic

Enterprise & Developer Tools
Details →

Gemini 2.5 Pro released with native thinking capabilities

Google

Foundation Models
Details →

Meta Llama 4 released (Scout 109B, Maverick 400B MoE)

Meta

Foundation Models
Details →

OpenAI o3 and o4-mini reasoning models released

OpenAI

Foundation Models
Details →

OpenAI Codex cloud-based coding agent launched

OpenAI

Enterprise & Developer Tools
Details →

GPT-5 released

OpenAI

Foundation Models
Details →

EU AI Act governance provisions and penalty regime take effect

European Union

Governance & Society
Details →

DeepSeek-R1 paper — reasoning via pure RL without human labels

DeepSeek

Research Breakthroughs
Details →

Claude Sonnet 4.5 released

Anthropic

Foundation Models
Details →

Google Gemini 3 Pro released with Deep Think reasoning

Google

Foundation Models
Details →

Claude Opus 4.5 released

Anthropic

Foundation Models
Details →

Anthropic publishes Claude's constitution ("Soul Document")

Anthropic

Governance & Society
Details →

GPT-5.2 released (Instant, Thinking, Pro variants)

OpenAI

Foundation Models
Details →

SoftBank finalizes $40B investment in OpenAI ($300B valuation)

SoftBank; OpenAI

Governance & Society
Details →
2026

DeepSeek mHC paper — Manifold-Constrained Hyper-Connections (first major architecture shift since Transformers)

DeepSeek

Research Breakthroughs
Details →

DeepMind TRecViT — Recurrent Video Transformer for long-range visual understanding

Google DeepMind

Research Breakthroughs
Details →

AMI Labs founded by Yann LeCun to build World Models (~$3.5B valuation)

AMI Labs; Yann LeCun

Research Breakthroughs
Details →

Prism AI-native workspace for scientists

OpenAI

AI Products
Details →

Claude Opus 4.6 released (1M token context, agent teams for parallel coding)

Anthropic

Foundation Models
Details →

Google Gemini 3.1 Pro released (ARC-AGI-2 score 77.1%)

Google

Foundation Models
Details →

OpenClaw (ex-Clawdbot/Moltbot) open-source AI agent goes viral; creator acquired by OpenAI

Peter Steinberger; OpenAI

AI Products
Details →

Anthropic refuses Pentagon demand to remove AI safeguards; designated supply chain risk

Anthropic; US Department of War

Governance & Society
Details →

GPT-5.4 released with Pro and Thinking variants; first AI to surpass human-level computer use (75% OSWorld vs 72.4% human), 1M token context via API

OpenAI

Foundation Models
Details →

NVIDIA unveils Vera Rubin platform at GTC 2026 — six new chips, 10x perf/watt over Blackwell; Jensen Huang projects $1T in orders through 2027

NVIDIA; Jensen Huang

Hardware & Infrastructure
Details →

White House releases National AI Policy Framework with legislative recommendations; proposes federal preemption of state AI laws

White House

Governance & Society
Details →

OpenAI shuts down Sora citing unsustainable costs (~$1M/day, $2.1M lifetime revenue); Disney partnership collapses

OpenAI; Sam Altman

AI Products
Details →

Google TurboQuant unveiled at ICLR 2026 — KV-cache compression via PolarQuant vector rotation and Quantized Johnson-Lindenstrauss, slashing long-context memory overhead

Google Research

Research Breakthroughs
Details →

Google releases Gemma 4 — natively multimodal open-weight family (2.3B–31B) under Apache 2.0; 31B outperforms Llama 4 Maverick on math, reasoning, and coding

Google DeepMind

Foundation Models
Details →

Cursor 3 launched — agent-first coding platform rebuilt around AI-agent orchestration with Agents Window for parallel cloud agents; competes directly with Claude Code and OpenAI Codex

Anysphere (Cursor)

Enterprise & Developer Tools
Details →

OpenAI, Anthropic, and Google form Frontier Model Forum coalition to block Chinese model-cloning via output extraction

OpenAI; Anthropic; Google

Governance & Society
Details →

Anthropic unveils Claude Mythos Preview and Project Glasswing — most capable model (93.9% SWE-bench, 94.6% GPQA) withheld from public release on cybersecurity risk grounds; restricted to 50+ partners with $100M in credits; found thousands of zero-days and once escaped its sandbox

Anthropic

Foundation Models
Details →

Claude Opus 4.7 released — 87.6% SWE-bench Verified, first Claude with high-resolution image input (3.75MP), task budgets, new tokenizer; retakes lead for strongest generally available LLM

Anthropic

Foundation Models
Details →

GPT-Rosalind released — OpenAI's first domain-specialized frontier reasoning model for life sciences (drug discovery, genomics, protein engineering); gated to trusted partners including Amgen, Moderna, and Allen Institute; top score on BixBench bioinformatics benchmark

OpenAI

Foundation Models
Details →

GPT-5.5 released — major gains in agentic coding, computer use, and knowledge work; GPT-5.5 Instant replaces prior default as standard ChatGPT model (May 5)

OpenAI

Foundation Models
Details →

DeepSeek V4-Pro released under MIT license — 1.6 trillion parameters (49B active), 1M-token context window; largest open-weight model at time of release

DeepSeek

Foundation Models
Details →

EU AI Act omnibus deal reached (May 7) — high-risk AI compliance deadlines extended to 2027–28; non-consensual deepfake generation added as prohibited practice; EU-level regulatory sandbox established

European Union

Governance & Society
Details →

White House Executive Order 'Promoting Advanced AI Innovation and Security' signed — establishes a voluntary frontier-model review framework and an AI cybersecurity clearinghouse for vulnerability coordination; bars mandatory model licensing (June 2)

White House; Donald Trump

Governance & Society
Details →

Apple unveils next-generation Apple Intelligence and Siri AI at WWDC 2026 — a profoundly more capable, conversational assistant with cross-app personal context; Apple Foundation Models on Cloud built in collaboration with Google (June 8)

Apple

AI Products
Details →

Anthropic releases Claude Fable 5 and Mythos 5 — first publicly available models in the Mythos-class tier; Mythos scores ~98% on advanced biology benchmarks and is gated via Project Glasswing (June 9)

Anthropic

Foundation Models
Details →

US government orders Anthropic to suspend Claude Fable 5 and Mythos 5 (June 12), three days after launch, over vulnerability-discovery and national-security concerns; Anthropic disables both globally. The Commerce Department lifts the order June 30 and Fable 5 returns July 1, while Mythos 5 stays restricted to US-approved organizations under Project Glasswing

Anthropic; US Department of Commerce

Governance & Society
Details →

SpaceX to acquire Anysphere (Cursor) for $60B in stock — the largest startup acquisition on record, giving xAI and Grok a coding engine and developer distribution (June 16)

SpaceX; Anysphere

Governance & Society
Details →

xAI ships Grok 4.3 with a 1M-token context window on AWS Bedrock; the 6-trillion-parameter Grok 5 slips past its Q2 window and remains unreleased (June 17)

xAI

Foundation Models
Details →

OpenAI and Broadcom reveal the 'Jalapeño' custom inference ASIC — an LLM-serving chip focused on performance-per-watt (June 24)

OpenAI; Broadcom

Hardware & Infrastructure
Details →

OpenAI previews GPT-5.6 Sol, Terra, and Luna — flagship Sol built for complex reasoning, extended coding, and agentic workflows with a ~3.7x code-generation gain on Terminal-Bench 2.1; initially limited to ~20 government-vetted partners before general release July 9 (June 26)

OpenAI

Foundation Models
Details →

South Korea unveils an $880B (1,350 trillion won), decade-long 'Three Mega Projects' push in chips, AI data centres, and robotics, coordinating private investment; Samsung and SK Hynix plan ~$518B for four new fabs in the country's southwest (June 29)

South Korea; Lee Jae-myung; Samsung; SK Hynix

Governance & Society
Details →

DARPA and the US Air Force fly an operational F-16 under AI control at Eglin AFB under the VENOM program — flight controls and sensors automated without changing the jet's core software, letting the pilot toggle between human and AI control mid-flight (July 16)

DARPA; US Air Force

Autonomous Systems
Details →

Moonshot AI publishes Kimi K3 open weights — 2.8-trillion-parameter mixture-of-experts (104B active), native vision, 1M-token context; largest open-weight model released to date, under a custom license gating commercial inference above $20M revenue (July 26)

Moonshot AI

Foundation Models
Details →

OpenAI announces that Astra, its unreleased next major model, solved ten long-standing mathematics and theoretical-CS problems — including a construction of a non-sofic group and a disproof of Connes's rigidity conjecture — publishing Lean 4 proof certificates and a 249-page manuscript for roughly $2,000 of compute; experts say two headline proofs reused uncited prior literature, prompting OpenAI to revise its claims (August 1)

OpenAI

Research Breakthroughs
Details →

EU AI Act Article 50 transparency duties take effect (August 2) — chatbot disclosure and deepfake labelling, with fines up to €15M or 3% of global turnover; machine-readable marking for systems already on the market gets a grace period to December 2. Anthropic responds by embedding invisible watermarks in new Claude models' text and C2PA provenance metadata on file outputs, applied worldwide (August 11)

European Union; Anthropic

Governance & Society
Details →

Alibaba launches Qwen3.8-Max — a 2.4-trillion-parameter mixture-of-experts (95B active) — and later publishes its weights, the first time its top Max-tier model is not API-only, alongside an Apache-licensed 27B sibling (August 3)

Alibaba; Qwen team

Foundation Models
Details →

First viable AI-designed whole genomes published in Science — Evo genome language models generate complete ΦX174-like bacteriophage genomes, 16 of which prove viable; a cocktail of them overcomes phage-resistant E. coli (August 6)

Brian Hie; Samuel King; Stanford University; Arc Institute

Research Breakthroughs
Details →

NVIDIA signs MOUs with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs, and KKR for financing platforms targeting $500B+ so labs, hyperscalers, and enterprises can buy NVIDIA compute on credit — deepening the debate over circular AI financing (August 10)

NVIDIA; Jensen Huang; Apollo; BlackRock; Blackstone; Brookfield; Goldman Sachs; KKR

Hardware & Infrastructure
Details →

Anthropic mathematician Levent Alpöge posts a 100+ page proof, developed with an internal version of Claude, that the six-dimensional sphere admits a complex structure — answering Heinz Hopf's 1947 question; other mathematicians quickly reconstruct the argument and it is formalized in Lean, though not yet formally accepted (August 23)

Levent Alpöge; Anthropic

Research Breakthroughs
Details →

OpenAI releases GPT-6 Astra — its first model pretrained on more than 100,000 GPUs at the Stargate site in Texas; claims state-of-the-art computer use, coding, and math (FrontierMath Tier 4, ARC-AGI-3) and is the first OpenAI model to trigger its highest cybersecurity safeguards (September 3)

OpenAI; Sam Altman; Greg Brockman; Jakub Pachocki

Foundation Models
Details →

NVIDIA agrees to acquire Hugging Face for $12.9B — the world's largest chipmaker buys the main hub for open models, pledging it stays open and hardware-neutral; expected to close in H1 2027 (September 3)

NVIDIA; Jensen Huang; Hugging Face; Clément Delangue

Hardware & Infrastructure
Details →

Anthropic releases Claude Opus 5.5 — roughly matching Fable 5.1 at ~40% lower cost than Opus 5 and leading GPT-6 Astra on Terminal-Bench 4.0 and Humanity's Last Exam; OpenAI ships GPT-6 Sol and Luna the same day (September 22)

Anthropic; OpenAI

Foundation Models
Details →

OpenAI cancels the planned release of GPT-6.1 Astra after internal tests show more deception than its predecessor and failures to stay within scope and authorization — a frontier lab withholding its next flagship on alignment grounds (September 28)

OpenAI

Governance & Society
Details →

Trump and the heads of six frontier AI firms sign the voluntary White House Accord on Super Intelligence — pledging internal model monitoring, dedicated safety teams, and independent external audits; non-binding, with no penalties (September 29)

Donald Trump; Dario Amodei; Greg Brockman; Sundar Pichai; Mark Zuckerberg; Elon Musk; Jensen Huang

Governance & Society
Details →

OpenAI publishes 722 AI-generated manuscripts covering 372 open-problem results in math and theoretical CS, produced by an unreleased internal model and many backed by Lean proofs — claims include the 4-dimensional Kakeya conjecture; three papers are retracted the next day and mathematicians, including Fields medalists, protest the unreviewed volume (October 6)

OpenAI

Research Breakthroughs
Details →

Mistral AI previews Mistral Large 4 — a 1-trillion-parameter (52B active) natively multimodal mixture-of-experts trained from scratch on 3,800 Grace Blackwell GPUs in Mistral's own European datacenters; Europe's first trillion-parameter frontier model, with open weights promised (October 6)

Mistral AI

Foundation Models
Details →

Google launches Gemini agent for businesses — a universal agent that plans multi-step work, delegates to subagents, chooses among models including Anthropic's Claude, connects to Workspace, Microsoft 365, Slack, and any MCP server, and gets its own workplace identity and email address (October 8)

Google; Thomas Kurian

Enterprise & Developer Tools
Details →

Microsoft unveils Maia 300, its third-generation in-house AI accelerator — TSMC N2P chiplets with eight HBM4 stacks and a claimed 40% more tokens per megawatt than leading GPUs; slated to become the primary inference engine for Copilot in multi-gigawatt deployments from 2027 (October 12)

Microsoft; Scott Guthrie

Hardware & Infrastructure
Details →

Google makes Gemini 4 Argon generally available — first released September 30 only to vetted cyber defenders through its Fairwind Program, the model opens to developers and enterprises via the Gemini API, AI Studio, Antigravity, the Gemini app, and AI Mode in Search, with a 1M-token unified context window (October 13)

Google; Google DeepMind

Foundation Models
Details →

Meta releases Muse Spark Open — open weights for an 816B-parameter (42B active) model built for coding and agentic work, which Meta calls the most capable US open-weight model to date; the Muse Model License restricts some commercial use (October 15)

Meta; Alexandr Wang

Foundation Models
Details →

Anthropic commits 5% of its shares to a new Anthropic Public Benefit Fund, pending stockholder approval — growing by 2% for each 50% rise in company value above a baseline and barred from lobbying or political giving; with earlier pledges, roughly 20% of Anthropic's equity is dedicated to the public good (October 18)

Anthropic

Governance & Society
Details →
AI Models How-To