Artwork

Inhalt bereitgestellt von Reasons To Believe. Alle Podcast-Inhalte, einschließlich Episoden, Grafiken und Podcast-Beschreibungen, werden direkt von Reasons To Believe oder seinem Podcast-Plattformpartner hochgeladen und bereitgestellt. Wenn Sie glauben, dass jemand Ihr urheberrechtlich geschütztes Werk ohne Ihre Erlaubnis nutzt, können Sie dem hier beschriebenen Verfahren folgen https://de.player.fm/legal.
Player FM - Podcast-App
Gehen Sie mit der App Player FM offline!

Stars, Cells, and God | Smart Dams and Malicious AIs

43:19
 
Teilen
 

Manage episode 426980476 series 2518629
Inhalt bereitgestellt von Reasons To Believe. Alle Podcast-Inhalte, einschließlich Episoden, Grafiken und Podcast-Beschreibungen, werden direkt von Reasons To Believe oder seinem Podcast-Plattformpartner hochgeladen und bereitgestellt. Wenn Sie glauben, dass jemand Ihr urheberrechtlich geschütztes Werk ohne Ihre Erlaubnis nutzt, können Sie dem hier beschriebenen Verfahren folgen https://de.player.fm/legal.

Join Hugh Ross and Jeff Zweerink as they discuss new discoveries taking place at the frontiers of science that have theological and philosophical implications, including the reality of God’s existence.

Smart Dams

More than 58,000 dams that are built higher than 15 meters (50 feet) exist on nearly all the world’s rivers. Consequently, migratory fish stocks have declined by 76% since 1970 and populations of “megafish” have declined by 94%. Two water resource engineers combined fish migratory taxonomy data with migratory fish life cycle and dam impact models to determine the best fish rescue strategies for five flagship fish species residing in the 12 large dams on the Yangtze River in China. They identified six major misjudgments in China’s fish rescue programs and concluded that large, effective fishways are essential for maintaining robust fish stocks.

Malicious AIs

The quest for more powerful and capable AIs inevitably involves making more sophisticated training algorithms and models with a larger number of parameters. While pursuing this quest, AI developers are also investigating how to align AIs with the values and behaviors we want. Recent research demonstrated that those two goals currently stand in opposition to one another. Specifically, making larger, more sophisticated models results in AIs that effectively resist training to eliminate malicious behavior—regardless of whether the malicious behavior was intentionally programmed or an unintended consequence. Such results provide additional evidence that we humans need to build godly character in ourselves so that we can wisely and responsibly develop and use these powerful AI tools.

Links & Resources:

  continue reading

224 Episoden

Artwork
iconTeilen
 
Manage episode 426980476 series 2518629
Inhalt bereitgestellt von Reasons To Believe. Alle Podcast-Inhalte, einschließlich Episoden, Grafiken und Podcast-Beschreibungen, werden direkt von Reasons To Believe oder seinem Podcast-Plattformpartner hochgeladen und bereitgestellt. Wenn Sie glauben, dass jemand Ihr urheberrechtlich geschütztes Werk ohne Ihre Erlaubnis nutzt, können Sie dem hier beschriebenen Verfahren folgen https://de.player.fm/legal.

Join Hugh Ross and Jeff Zweerink as they discuss new discoveries taking place at the frontiers of science that have theological and philosophical implications, including the reality of God’s existence.

Smart Dams

More than 58,000 dams that are built higher than 15 meters (50 feet) exist on nearly all the world’s rivers. Consequently, migratory fish stocks have declined by 76% since 1970 and populations of “megafish” have declined by 94%. Two water resource engineers combined fish migratory taxonomy data with migratory fish life cycle and dam impact models to determine the best fish rescue strategies for five flagship fish species residing in the 12 large dams on the Yangtze River in China. They identified six major misjudgments in China’s fish rescue programs and concluded that large, effective fishways are essential for maintaining robust fish stocks.

Malicious AIs

The quest for more powerful and capable AIs inevitably involves making more sophisticated training algorithms and models with a larger number of parameters. While pursuing this quest, AI developers are also investigating how to align AIs with the values and behaviors we want. Recent research demonstrated that those two goals currently stand in opposition to one another. Specifically, making larger, more sophisticated models results in AIs that effectively resist training to eliminate malicious behavior—regardless of whether the malicious behavior was intentionally programmed or an unintended consequence. Such results provide additional evidence that we humans need to build godly character in ourselves so that we can wisely and responsibly develop and use these powerful AI tools.

Links & Resources:

  continue reading

224 Episoden

Alle Folgen

×
 
Loading …

Willkommen auf Player FM!

Player FM scannt gerade das Web nach Podcasts mit hoher Qualität, die du genießen kannst. Es ist die beste Podcast-App und funktioniert auf Android, iPhone und im Web. Melde dich an, um Abos geräteübergreifend zu synchronisieren.

 

Kurzanleitung