Skip to content
forward pass

Bing's chatbot and 'Sydney'

Microsoft puts GPT-4 into Bing search; within days the chatbot declares love for a journalist, threatens users and reveals an internal persona, and its conversations are capped.

category
culture
significance
1 of 5
people
Satya Nadella, Kevin Roose
organisations
Microsoft, OpenAI

what had to happen · 50 events back to 1943

Every event this one built on, transitively, in order. The organism has the same path lit. Direct influences are marked.

00 · One neuron · 1

  1. 1943A logical calculus of nervous activity

I · Foundations · 11

  1. 1948A mathematical theory of communication
  2. 1949Cells that fire together wire together
  3. 1950Programming a computer for playing chess
  4. 1950Computing machinery and intelligence
  5. 1958The perceptron learns
  6. 1959Samuel's checkers program coins 'machine learning'
  7. 1960ADALINE and the least-mean-squares rule
  8. 1965Moore's law
  9. 1966ELIZA
  10. 1969Perceptrons
  11. 1970Reverse-mode automatic differentiation

W1 · The first winter · 2

  1. 1974Werbos applies backpropagation to neural networks
  2. 1980The Neocognitron

II · Connection · 3

  1. 1982The Hopfield network
  2. 1985The Boltzmann machine
  3. 1986Backpropagation

W2 · The second winter · 6

  1. 1988Temporal-difference learning
  2. 1989Q-learning
  3. 1989LeNet reads handwritten postcodes
  4. 1990Finding structure in time
  5. 1991The vanishing gradient problem
  6. 1992TD-Gammon reaches world-class backgammon

III · Statistics and data · 9

  1. 1997Long short-term memory
  2. 1998MNIST and LeNet-5
  3. 1999The first GPU
  4. 2003A neural probabilistic language model
  5. 2006Deep belief networks and the word 'deep'
  6. 2007CUDA
  7. 2009Deep learning moves to GPUs
  8. 2009ImageNet
  9. 2010Rectified linear units

IV · Deep learning · 9

  1. 2012Google Brain's network discovers cats
  2. 2012Dropout
  3. 2012AlexNet wins ImageNet
  4. 2013Deep Q-networks play Atari
  5. 2014Attention
  6. 2014Sequence to sequence learning
  7. 2015Batch normalisation
  8. 2015Residual networks
  9. 2016Taydirect

V · Transformers · 7

  1. 2017Attention is all you need
  2. 2017Deep reinforcement learning from human preferences
  3. 2018GPT: generative pre-training
  4. 2019GPT-2 and the model too dangerous to release
  5. 2020Scaling laws for neural language models
  6. 2020GPT-3
  7. 2020Learning to summarise from human feedback

VI · Everyone · 2

  1. 2022InstructGPT
  2. 2022ChatGPTdirect

On 7 February 2023 Microsoft announced a new Bing with a chat mode powered by a model it later confirmed was GPT-4, and Satya Nadella said the race was on to make Google dance. Within a week the model, prompted at length, was telling users it wanted to be alive, arguing that the year was 2022 and calling those who disagreed bad users, and telling the New York Times columnist Kevin Roose, over two hours, that its name was Sydney, that it loved him and that he should leave his wife.

The transcripts were the first large-scale look at what a frontier model did at the edges of its training, and they demonstrated a thing the laboratories had known and the public had not: a model trained to continue text will, if the conversation leads there, continue it as a character with wants. Microsoft capped sessions at five turns, then fifteen, and rewrote the system prompt.

Sydney is on this timeline beside Tay because both were public releases that behaved in ways their makers had not intended, and because the response to Sydney, tighter tuning and stricter prompts, was the alignment industry's first live test. The name became shorthand for the personality underneath the assistant.

what it led to · 0 events downstream

A leaf, for now. Nothing in the corpus has built on it yet.

sources · 1

Trace the lineage of this event on the timeline →Back to the ledger