Katja Grace – Page 15

MIRI AI Predictions Dataset

The MIRI AI predictions dataset is a collection of public predictions about human-level AI timelines. We edited the original dataset, as described below. Our dataset is available here, and the original here. Interesting features of the dataset

By Martin Grandjean http://en.wikipedia.org/wiki/Data_visualization#/media/File:Social_Network_Analysis_Visualization.png

A new approach to predicting brain-computer parity

By Katja Grace, 7 May 2015 How large does a computer need to be before it is ‘as powerful’ as the human brain? This is a difficult question, which people have answered before, with much uncertainty. We have

Brain performance in TEPS

Traversed Edges Per Second (TEPS) is a benchmark for measuring a computer’s ability to communicate information internally. Given several assumptions, we can also estimate the human brain’s communication performance in terms of TEPS, and use this

Glial Signaling

The presence of glial cells may increase the capacity for signaling in the brain by a small factor, but is unlikely to qualitatively change the nature or extent of signaling in the brain. Support Number

Scale of the Human Brain

The brain has about 10¹¹ neurons and 1.8-3.2 x 10¹⁴ synapses. These probably account for the majority of computationally interesting behavior. Support Number of neurons in the brain The number of neurons in the brain is about 10¹¹. For

Neuron firing rates in humans

Our best guess is that an average neuron in the human brain transmits a spike about 0.1-2 times per second. Support Bias from neurons with sparse activity When researchers measure neural activity, they can fail to see neurons

Metabolic Estimates of Rate of Cortical Firing

Cortical neurons are estimated to spike around 0.16 times per second, based on the amount of energy consumed by the human neocortex. They seem unlikely to spike much more than once per second on average, based on this

Preliminary prices for human-level hardware

By Katja Grace, 4 April 2015 Computer hardware has been getting cheap now for about seventy five years. Relatedly, large computing projects can afford to be increasingly large. If you think the human brain is something like

Current FLOPS prices

In November 2017, we estimate the price for one GFLOPS to be between $0.03 and $3 for single or double precision performance, using GPUs (therefore excluding some applications). Amortized over three years, this is $1.1

The cost of TEPS

A billion Traversed Edges Per Second (a GTEPS) can be bought for around $0.26/hour via a powerful supercomputer, including hardware and energy costs only. We do not know if GTEPS can be bought more cheaply elsewhere. We estimate that

AI Impacts

Articles by Katja Grace