113
个编辑
更改
跳到导航
跳到搜索
第180行：
第180行： + + − In a hard takeoff scenario, an AGI rapidly self-improves, "taking control" of the world (perhaps in a matter of hours), too quickly for significant human-initiated error correction or for a gradual tuning of the AGI's goals. In a soft takeoff scenario, AGI still becomes far more powerful than humanity, but at a human-like pace (perhaps on the order of decades), on a timescale where ongoing human interaction and correction can effectively steer the AGI's development.<ref>Bugaj, Stephan Vladimir, and Ben Goertzel. "Five ethical imperatives and their implications for human-AGI interaction." Dynamical Psychology (2007).</ref><ref>Sotala, Kaj, and Roman V. Yampolskiy. "Responses to catastrophic AGI risk: a survey." Physica Scripta 90.1 (2014): 018001.</ref>+ − 在硬起飞的情况下，通用人工智能 AGI 迅速自我完善并“掌控”世界（也许在几个小时内）。速度太快，以致于无法进行重大的人为纠错或对 AGI 的目标进行逐步调整。在软起飞的情况下，虽然 AGI 仍然比人类强大得多，但却以一种类似人类的速度进步（也许是几十年的数量级），持续的人类互动和修正可以有效地引导AGI 的发展。<ref>Bugaj, Stephan Vladimir, and Ben Goertzel. "Five ethical imperatives and their implications for human-AGI interaction." Dynamical Psychology (2007).</ref><ref>Sotala, Kaj, and Roman V. Yampolskiy. "Responses to catastrophic AGI risk: a survey." Physica Scripta 90.1 (2014): 018001.</ref>+ − + − [[Ramez Naam]] argues against a hard takeoff. He has pointed that we already see recursive self-improvement by superintelligences, such as corporations. [[Intel]], for example, has "the collective brainpower of tens of thousands of humans and probably millions of CPU cores to... design better CPUs!" However, this has not led to a hard takeoff; rather, it has led to a soft takeoff in the form of [[Moore's law]].<ref name=Naam2014Further>{{cite web|last=Naam|first=Ramez|title=The Singularity Is Further Than It Appears|url=http://www.antipope.org/charlie/blog-static/2014/02/the-singularity-is-further-tha.html|accessdate=16 May 2014|year=2014}}</ref> Naam further points out that the computational complexity of higher intelligence may be much greater than linear, such that "creating a mind of intelligence 2 is probably ''more'' than twice as hard as creating a mind of intelligence 1."<ref name=Naam2014Ascend>{{cite web|last=Naam|first=Ramez|title=Why AIs Won't Ascend in the Blink of an Eye - Some Math|url=http://www.antipope.org/charlie/blog-static/2014/02/why-ais-wont-ascend-in-blink-of-an-eye.html|accessdate=16 May 2014|year=2014}}</ref> − − − Ramez Naam 反对硬起飞。他指出，我们已经看到企业等超级智能体的递归式地自我完善。例如，Intel拥有“数万人的集体脑力，可能还有数百万个CPU核心……来设计更好的CPU！”然而，这并没有导致硬起飞；相反，它以摩尔定律的形式实现了软起飞<ref name=Naam2014Further>{{cite web|last=Naam|first=Ramez|title=The Singularity Is Further Than It Appears|url=http://www.antipope.org/charlie/blog-static/2014/02/the-singularity-is-further-tha.html|accessdate=16 May 2014|year=2014}}</ref> 。Naam进一步指出，高等智能的计算复杂度可能比线性大得多，因此，“创建智能系统2的难度可能两倍于是创建智能系统1的两倍多。”<ref name=Naam2014Ascend>{{cite web|last=Naam|first=Ramez|title=Why AIs Won't Ascend in the Blink of an Eye - Some Math|url=http://www.antipope.org/charlie/blog-static/2014/02/why-ais-wont-ascend-in-blink-of-an-eye.html|accessdate=16 May 2014|year=2014}}</ref> − − [[J. Storrs Hall]] believes that "many of the more commonly seen scenarios for overnight hard takeoff are circular – they seem to assume hyperhuman capabilities at the ''starting point'' of the self-improvement process" in order for an AI to be able to make the dramatic, domain-general improvements required for takeoff. Hall suggests that rather than recursively self-improving its hardware, software, and infrastructure all on its own, a fledgling AI would be better off specializing in one area where it was most effective and then buying the remaining components on the marketplace, because the quality of products on the marketplace continually improves, and the AI would have a hard time keeping up with the cutting-edge technology used by the rest of the world.<ref name=Hall2008>{{cite journal|last=Hall|first=J. Storrs|title=Engineering Utopia|journal=Artificial General Intelligence, 2008: Proceedings of the First AGI Conference|date=2008|pages=460–467|url=http://www.agiri.org/takeoff_hall.pdf|accessdate=16 May 2014}}</ref> − − J.Storrs Hall 认为，“许多常见的一夜之间出现的硬起飞场景都是循环论证——它们似乎在自我提升过程的起点上假设了超人类的能力”，以便人工智能能够实现起飞所需的戏剧性的、领域通用的改进。Hall 认为，一个初出茅庐的人工智能与其靠自己不断地自我改进硬件、软件和基础设施，不如专注于一个它最有效的领域，然后在市场上购买剩余的组件，因为市场上产品的质量不断提高，人工智能将很难跟上世界其他地方使用的尖端技术。<ref name=Hall2008>{{cite journal|last=Hall|first=J. Storrs|title=Engineering Utopia|journal=Artificial General Intelligence, 2008: Proceedings of the First AGI Conference|date=2008|pages=460–467|url=http://www.agiri.org/takeoff_hall.pdf|accessdate=16 May 2014}}</ref> − − − [[Ben Goertzel]] agrees with Hall's suggestion that a new human-level AI would do well to use its intelligence to accumulate wealth. The AI's talents might inspire companies and governments to disperse its software throughout society. Goertzel is skeptical of a hard five minute takeoff but speculates that a takeoff from human to superhuman level on the order of five years is reasonable. Goerzel refers to this scenario as a "semihard takeoff".<ref name="Goertzel2014">{{cite news|last1=Goertzel|first1=Ben|title=Superintelligence — Semi-hard Takeoff Scenarios|url=http://hplusmagazine.com/2014/09/26/superintelligence-semi-hard-takeoff-scenarios/|accessdate=25 October 2014|agency=h+ Magazine|date=26 Sep 2014}}</ref> − − Ben Goertzel 同意 Hall 的意见，即一种新的人类级别的人工智能将很好地利用其智能来积累财富。人工智能的天赋可能会激励公司和政府将其软件分散到整个社会。Goertzel 对5分钟的硬起飞持怀疑态度，但他推测从人类到超人的水平，以5年的速度起飞是合理的。Goerzel将这种情况称为“半硬起飞semihard takeoff”。<ref name="Goertzel2014">{{cite news|last1=Goertzel|first1=Ben|title=Superintelligence — Semi-hard Takeoff Scenarios|url=http://hplusmagazine.com/2014/09/26/superintelligence-semi-hard-takeoff-scenarios/|accessdate=25 October 2014|agency=h+ Magazine|date=26 Sep 2014}}</ref> − − [[Max More]] disagrees, arguing that if there were only a few superfast human-level AIs, that they would not radically change the world, as they would still depend on other people to get things done and would still have human cognitive constraints. Even if all superfast AIs worked on intelligence augmentation, it is unclear why they would do better in a discontinuous way than existing human cognitive scientists at producing super-human intelligence, although the rate of progress would increase. More further argues that a superintelligence would not transform the world overnight: a superintelligence would need to engage with existing, slow human systems to accomplish physical impacts on the world. "The need for collaboration, for organization, and for putting ideas into physical changes will ensure that all the old rules are not thrown out overnight or even within years."<ref name=More>{{cite web|last1=More|first1=Max|title=Singularity Meets Economy|url=http://hanson.gmu.edu/vc.html#more|accessdate=10 November 2014}}</ref> − − Max More 不同意这一观点，他认为，如果只有少数超高速的人类水平的人工智能，它们不会从根本上改变世界，因为它们仍将依赖人来完成任务，并且仍然会受到人类认知的限制。即使所有的超高速人工智能都致力于智能增强，但目前还不清楚为什么它们在产生超人类智能方面比现有的人类认知科学家做得更好，尽管进展速度会加快。更进一步指出，超级智能不会在一夜之间改变世界：超级智能需要与现有的、缓慢的人类系统进行接触，以完成对世界的物理影响。”合作、组织和将想法付诸实际变革的需要将确保所有旧规则不会在一夜之间甚至几年内被抛弃。”<ref name=More>{{cite web|last1=More|first1=Max|title=Singularity Meets Economy|url=http://hanson.gmu.edu/vc.html#more|accessdate=10 November 2014}}</ref>
→硬起飞与软起飞
[[文件：在这个示例递归自我改进场景中，修改人工智能体系结构的人可以每三年将其性能提高一倍，例如，30代人，然后用尽所有可行的改进（左）。相反，如果人工智能足够聪明，能够像人类研究人员那样修改自己的架构，那么每一代人完成一次重新设计所需的时间将减半，并且它在6年内将所有30代可行的代都推进（右图）。<ref name="yudkowsky-global-risk">[[Eliezer Yudkowsky]]. "Artificial intelligence as a positive and negative factor in global risk." Global catastrophic risks (2008).</ref>]]
[[文件：在这个示例递归自我改进场景中，修改人工智能体系结构的人可以每三年将其性能提高一倍，例如，30代人，然后用尽所有可行的改进（左）。相反，如果人工智能足够聪明，能够像人类研究人员那样修改自己的架构，那么每一代人完成一次重新设计所需的时间将减半，并且它在6年内将所有30代可行的代都推进（右图）。<ref name="yudkowsky-global-risk">[[Eliezer Yudkowsky]]. "Artificial intelligence as a positive and negative factor in global risk." Global catastrophic risks (2008).</ref>]]
在硬起飞的情况下，通用人工智能AGI迅速自我完善并“掌控”世界（也许在几个小时内）。速度太快，以致于无法进行重大的人为纠错或对 AGI 的目标进行逐步调整。在软起飞的情况下，虽然AGI仍然比人类强大得多，但却以一种类似人类的速度进步（也许是几十年的数量级），持续的人类互动和修正可以有效地引导AGI的发展。<ref>Bugaj, Stephan Vladimir, and Ben Goertzel. "Five ethical imperatives and their implications for human-AGI interaction." Dynamical Psychology (2007).</ref><ref>Sotala, Kaj, and Roman V. Yampolskiy. "Responses to catastrophic AGI risk: a survey." Physica Scripta 90.1 (2014): 018001.</ref>
Ramez Naam反对硬起飞。他指出，我们已经看到企业等超级智能体的递归式地自我完善。例如，Intel拥有“数万人的集体脑力，可能还有数百万个CPU核心……来设计更好的CPU！”然而，这并没有导致硬起飞；相反，它以摩尔定律的形式实现了软起飞。<ref name=Naam2014Further>{{cite web|last=Naam|first=Ramez|title=The Singularity Is Further Than It Appears|url=http://www.antipope.org/charlie/blog-static/2014/02/the-singularity-is-further-tha.html|accessdate=16 May 2014|year=2014}}</ref>Naam进一步指出，高等智能的计算复杂度可能比线性大得多，因此，“创建智能系统2的难度可能两倍于是创建智能系统1的两倍多。”<ref name=Naam2014Ascend>{{cite web|last=Naam|first=Ramez|title=Why AIs Won't Ascend in the Blink of an Eye - Some Math|url=http://www.antipope.org/charlie/blog-static/2014/02/why-ais-wont-ascend-in-blink-of-an-eye.html|accessdate=16 May 2014|year=2014}}</ref>
J.Storrs Hall认为，“许多常见的一夜之间出现的硬起飞场景都是循环论证——它们似乎在自我提升过程的起点上假设了超人类的能力”，以便人工智能能够实现起飞所需的戏剧性的、领域通用的改进。Hall认为，一个初出茅庐的人工智能与其靠自己不断地自我改进硬件、软件和基础设施，不如专注于一个它最有效的领域，然后在市场上购买剩余的组件，因为市场上产品的质量不断提高，人工智能将很难跟上世界其他地方使用的尖端技术。<ref name=Hall2008>{{cite journal|last=Hall|first=J. Storrs|title=Engineering Utopia|journal=Artificial General Intelligence, 2008: Proceedings of the First AGI Conference|date=2008|pages=460–467|url=http://www.agiri.org/takeoff_hall.pdf|accessdate=16 May 2014}}</ref>
Ben Goertzel同意Hall的意见，即一种新的人类级别的人工智能将很好地利用其智能来积累财富。人工智能的天赋可能会激励公司和政府将其软件分散到整个社会。Goertzel对5分钟的硬起飞持怀疑态度，但他推测从人类到超人的水平，以5年的速度起飞是合理的。Goerzel将这种情况称为“半硬起飞 semihard takeoff”。<ref name="Goertzel2014">{{cite news|last1=Goertzel|first1=Ben|title=Superintelligence — Semi-hard Takeoff Scenarios|url=http://hplusmagazine.com/2014/09/26/superintelligence-semi-hard-takeoff-scenarios/|accessdate=25 October 2014|agency=h+ Magazine|date=26 Sep 2014}}</ref>
Max More不同意这一观点，他认为，如果只有少数超高速的人类水平的人工智能，它们不会从根本上改变世界，因为它们仍将依赖人来完成任务，并且仍然会受到人类认知的限制。即使所有的超高速人工智能都致力于智能增强，但目前还不清楚为什么它们在产生超人类智能方面比现有的人类认知科学家做得更好，尽管进展速度会加快。更进一步指出，超级智能不会在一夜之间改变世界：超级智能需要与现有的、缓慢的人类系统进行接触，以完成对世界的物理影响。”合作、组织和将想法付诸实际变革的需要将确保所有旧规则不会在一夜之间甚至几年内被抛弃。”<ref name=More>{{cite web|last1=More|first1=Max|title=Singularity Meets Economy|url=http://hanson.gmu.edu/vc.html#more|accessdate=10 November 2014}}</ref>
==永生==
==永生==