Question

Will the US government deem that an AI model has been trained with at least 1e26 FLOP before 2026?

8 comments8

Total Forecasters29

Community Prediction94%

Make a Prediction

community: 94%

median 94.0%mean 90.0%

Authors:

odelaney PeterWildeford

Opened:

Closes:

Scheduled resolution:

Regulation of AI

Politics

Computing and Math

Artificial Intelligence

Information technology

United States

Will the FLOPs thresholds outlined in the US Executive Order on Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence be updated before the indicated year?

Will the estimated training FLOPs of any exfiltrated AI exceed the indicated amount before the indicated year?

When will a Chinese organization train a large model on GPUs that deliver more than 1500 TFLOPs@FP16?

2024-12

Comments

? comments

Authors:

odelaney PeterWildeford

Opened:

Closes:

Scheduled resolution:

Regulation of AI

Politics

Computing and Math

Artificial Intelligence

Information technology

United States

SoftBank to Invest $500M in OpenAI for the First Time—Is It Late to the Party?

Observer (NY)•Oct 1, 2024

Silicon Valley has a plan to save humanity: Just flip on the nuclear reactors

CNN•Oct 1, 2024

The AI boom turned Nvidia into a trillion-dollar giant. Its next big bet is even more complex.

Business Insider•Oct 1, 2024

Learn more about Metaculus NewsMatch

Will the FLOPs thresholds outlined in the US Executive Order on Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence be updated before the indicated year?

Will the estimated training FLOPs of any exfiltrated AI exceed the indicated amount before the indicated year?

When will a Chinese organization train a large model on GPUs that deliver more than 1500 TFLOPs@FP16?

2024-12

MetaculusM

🚀Metaculus Open-Source Rewrite is Live!

Will the US government deem that an AI model has been trained with at least 1e26 FLOP before 2026?

Make a Prediction

Will the FLOPs thresholds outlined in the US Executive Order on Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence be updated before the indicated year?

Will the estimated training FLOPs of any exfiltrated AI exceed the indicated amount before the indicated year?

When will a Chinese organization train a large model on GPUs that deliver more than 1500 TFLOPs@FP16?

Comments

Will the FLOPs thresholds outlined in the US Executive Order on Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence be updated before the indicated year?

Will the estimated training FLOPs of any exfiltrated AI exceed the indicated amount before the indicated year?

When will a Chinese organization train a large model on GPUs that deliver more than 1500 TFLOPs@FP16?

Will the FLOPs thresholds outlined in the US Executive Order on Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence be updated before the indicated year?

Will the estimated training FLOPs of any exfiltrated AI exceed the indicated amount before the indicated year?

When will a Chinese organization train a large model on GPUs that deliver more than 1500 TFLOPs@FP16?

Will the FLOPs thresholds outlined in the US Executive Order on Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence be updated before the indicated year?

Will the estimated training FLOPs of any exfiltrated AI exceed the indicated amount before the indicated year?

When will a Chinese organization train a large model on GPUs that deliver more than 1500 TFLOPs@FP16?