GenAI Chronicle
2010
2011
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2018

June

4 events

Editors' summary

Microsoft buys GitHub, Google sets out AI principles

On 4 June, Microsoft announced it would buy GitHub for $7.5 billion, taking hold of where developers gather. Parts of the open source world were wary.

On the 7th, Google published its AI Principles, stating it would not apply AI to weapons or to surveillance that violates international norms — its answer to April's protest over Project Maven.

On the 11th, OpenAI announced the first GPT: pre-train on a large body of text, then adapt per task. The point was that the groundwork could be laid from unlabelled text alone.

On the 25th it announced OpenAI Five, taking on five-on-five Dota 2.

This block is written by the editors. It is kept separate from the sourced record below.

Record4 events
  1. 4
    BusinessMicrosoft / Satya Nadella / Nat Friedman / Chris Wanstrath

    Microsoft acquires GitHub for $7.5 billion

    Microsoft announced an agreement to acquire GitHub, the world's largest software development platform, for $7.5 billion in stock; more than 28 million developers were then using it. CEO Satya Nadella said "Microsoft is a developer-first company, and by joining forces with GitHub we strengthen our commitment to developer freedom, openness and innovation." Microsoft's Nat Friedman became GitHub's CEO, and founder and former CEO Chris Wanstrath became a Microsoft technical fellow. The deal drew wide attention at the time as confirmation of Microsoft's turn away from hostility toward open source. It also gave Microsoft the largest collection of public source code in existence; GitHub Copilot, released in 2021, would be built on top of it.

  2. 7
    GovernanceGoogle / Sundar Pichai

    Google publishes AI Principles, ruling out weapons

    Sundar Pichai published "AI at Google: our principles," setting out seven principles for the company's AI work along with four areas it would not pursue. Those excluded "weapons or other technologies whose principal purpose or implementation is to cause or directly facilitate injury to people," technologies that "gather or use information for surveillance violating internationally accepted norms," and technologies "whose purpose contravenes widely accepted principles of international law and human rights." The document was Google's answer to the internal protest over Project Maven, and Pichai was explicit that the company would continue working with governments and militaries in other areas such as cybersecurity, training, and search and rescue. As one of the first instances of a major AI company publicly declaring what it would refuse to build, it became a template for comparable policies elsewhere.

  3. 11
    ModelOpenAI / Alec Radford

    Original GPT announced

    Alec Radford and colleagues at OpenAI announced the original GPT, later known as GPT-1, with the paper "Improving Language Understanding by Generative Pre-Training". Using a two-stage approach—unsupervised pre-training of a transformer on large amounts of text via next-word prediction, followed by fine-tuning on small supervised datasets—it achieved state-of-the-art results on diverse language understanding tasks including textual entailment, commonsense reasoning, and reading comprehension. The model and code were released. It was the first model to carry the GPT name.

  4. 25
    ResearchOpenAI

    OpenAI Five announced for 5v5 Dota 2

    OpenAI announced OpenAI Five, a team of five neural networks that plays 5v5 Dota 2. Trained via self-play without human data, it played the equivalent of 180 years of games against itself every day, using a scaled-up version of Proximal Policy Optimization (PPO) running on 256 GPUs and 128,000 CPU cores. Playing under restricted rules, it had started to defeat amateur human teams, and OpenAI set the goal of facing top professionals at The International in August.

↑ 2018