🤖70-Year-Old SEER Machine Still Beats Humans 59% of the Time
A 70-year-old machine still outsmarts humans in a game
TL;DR
SEER, a 70-year-old machine built at Bell Labs, still outsmarts humans in a game with a 59% win rate. Claude Shannon's machine, which played SEER, won 55.8% against it. The game tracks patterns and randomness, showing the enduring impact of early AI.
SEER, a machine built in the early 1950s at Bell Labs, still outsmarts humans in a game with a 59% win rate. The machine, designed by David Hagelbarger, tracks patterns and randomness to predict the next move. Claude Shannon's machine, which played SEER, won 55.8% against it in 2020 software replication. This shows the enduring impact of early AI on modern computing. The game uses a notebook to record every key press, from runs of five down to no context at all, making it a fascinating look into early machine learning.
Key Points
SEER was built in the early 1950s at Bell Labs by David Hagelbarger.
SEER tracks three yes-or-no facts about the last two rounds, winning or losing the play before last, playing the same or differently, and winning or losing the last play.
Claude Shannon's machine won 55.8% against SEER in 2020 software replication.
SEER's notebook records every key press, from runs of five down to no context at all.
The game uses runs of five and forgets nothing, showing the enduring impact of early AI.
Why It Matters
SEER, built in the early 1950s, still outsmarts humans in a game with a 59% win rate. This shows the enduring impact of early AI on modern computing. The game's use of pattern recognition and randomness is a precursor to today's machine learning techniques. Anyone interested in the history of AI and its evolution should take a look at SEER.
Comments
Be the first to comment
Enjoyed this article?
Get it daily. 7am. Free. Reads in 5 minutes.
Join 3,569 builders reading daily.