Sign up to save your podcasts
Or
Share The Only Source of Hope is Open
Share to email
Share to Facebook
Share to X
Share to email
Share to Facebook
Share to X
Or
The Only Source of Hope is Open
Summary:
The recent article "On the Opportunities and Risks of Foundation Models” from the Center for Research on Foundation Models and Stanford Institute for Human-Centered Artificial Intelligence" captures the risks but not the opportunities related to Foundation Models.
The significance of foundation models can be summarized by two words: emergence and homogenization.
The lack of diversity and redundancy in the AI ecosystem has already revealed the potentially dire consequences if the foundation model fails or is compromised.
Open-source competition is necessary to address ethical implications, environmental impact, and potential job loss related to foundation models.
Government incentives and private sector solutions like Stability AI are essential for growth in the market and the health, happiness, and prosperity of the citizenry.
Transcript:
Welcome to another episode of Courage Over Convention. In this episode, I will touch on the recent article "On the Opportunities and Risks of Foundation Models” from the Center for Research on Foundation Models and Stanford Institute for Human-Centered Artificial Intelligence." This article does a reasonable job of capturing the risks but not the opportunities related to Foundation Models. Unfortunately, the opportunities aren't worth contemplating if the risks are not adequately dealt with.
The rationale behind my assertion here is embodied in the paper’s most salient quote. The authors write: "The significance of foundation models can be summarized by two words: emergence and homogenization. Emergence means that the behavior of a system is implicitly induced rather than explicitly constructed; it is both the source of scientific excitement and anxiety about unanticipated consequences. Homogenization indicates the consolidation of methodologies for building machine learning systems across a wide range of applications; it provides strong leverage towards many tasks but also creates single points of failure."
So many applications now rely on ChatGPT. It is an enormous monopsonistic force. It has become, as Musk tweeted on Feb 17th, "a closed source, maximum-profit company effectively controlled by Microsoft."
This lack of diversity and redundancy in the AI ecosystem has already revealed the potentially dire consequences if the foundation model fails or is compromised.
While the CRFM and Stanford article discusses various aspects and implications of foundation models, it raises several problems, but does not go so far as to hint at a viable answer. This is, of course, to be expected in an academic paper. But still. C'mon. The answer is clear: open-source competition.
Firms like Stability AI are essential if we are to allow the private sector to drive growth in a market that is genuinely vital to national security, and the health, happiness, and prosperity of the citizenry. 1.73 billion dollars was allocated toward non-defense-related AI R&D from the U.S. Federal Budget last year, according to a recent release from Stanford. Meanwhile, Microsoft's revenue was a touch over 203 billion on the year. Clearly, collaboration and intervention via incentives, immediate intervention, is our only hope.
Keywords:
ChatGPT, LLMs, Stability AI, foundation models, risks, opportunities, emergence, homogenization, open-source, competition, AI ecosystem, diversity, redundancy, safety nets, federal government, private market, national security, citizenry.
Citations:
Center for Research on Foundation Models (CRFM) and Stanford Institute for Human-Centered Artificial Intelligence (HAI). (2021). On the Opportunities and Risks of Foundation Models. Retrieved from https://crfm.stanford.edu/assets/report.pdf
Stanford Institute for Human-Centered Artificial Intelligence. (2023, April 12). The State of AI in 2023: 14 Charts. Retrieved from https://hai.stanford.edu/news/2023-state-ai-14-charts.
---
Send in a voice message: https://podcasters.spotify.com/pod/show/couragevconvention/message
View all episodes
By Woody Wiegmann
5
66 ratings
Summary:
The recent article "On the Opportunities and Risks of Foundation Models” from the Center for Research on Foundation Models and Stanford Institute for Human-Centered Artificial Intelligence" captures the risks but not the opportunities related to Foundation Models.
The significance of foundation models can be summarized by two words: emergence and homogenization.
The lack of diversity and redundancy in the AI ecosystem has already revealed the potentially dire consequences if the foundation model fails or is compromised.
Open-source competition is necessary to address ethical implications, environmental impact, and potential job loss related to foundation models.
Government incentives and private sector solutions like Stability AI are essential for growth in the market and the health, happiness, and prosperity of the citizenry.
Transcript:
Welcome to another episode of Courage Over Convention. In this episode, I will touch on the recent article "On the Opportunities and Risks of Foundation Models” from the Center for Research on Foundation Models and Stanford Institute for Human-Centered Artificial Intelligence." This article does a reasonable job of capturing the risks but not the opportunities related to Foundation Models. Unfortunately, the opportunities aren't worth contemplating if the risks are not adequately dealt with.
The rationale behind my assertion here is embodied in the paper’s most salient quote. The authors write: "The significance of foundation models can be summarized by two words: emergence and homogenization. Emergence means that the behavior of a system is implicitly induced rather than explicitly constructed; it is both the source of scientific excitement and anxiety about unanticipated consequences. Homogenization indicates the consolidation of methodologies for building machine learning systems across a wide range of applications; it provides strong leverage towards many tasks but also creates single points of failure."
So many applications now rely on ChatGPT. It is an enormous monopsonistic force. It has become, as Musk tweeted on Feb 17th, "a closed source, maximum-profit company effectively controlled by Microsoft."
This lack of diversity and redundancy in the AI ecosystem has already revealed the potentially dire consequences if the foundation model fails or is compromised.
While the CRFM and Stanford article discusses various aspects and implications of foundation models, it raises several problems, but does not go so far as to hint at a viable answer. This is, of course, to be expected in an academic paper. But still. C'mon. The answer is clear: open-source competition.
Firms like Stability AI are essential if we are to allow the private sector to drive growth in a market that is genuinely vital to national security, and the health, happiness, and prosperity of the citizenry. 1.73 billion dollars was allocated toward non-defense-related AI R&D from the U.S. Federal Budget last year, according to a recent release from Stanford. Meanwhile, Microsoft's revenue was a touch over 203 billion on the year. Clearly, collaboration and intervention via incentives, immediate intervention, is our only hope.
Keywords:
ChatGPT, LLMs, Stability AI, foundation models, risks, opportunities, emergence, homogenization, open-source, competition, AI ecosystem, diversity, redundancy, safety nets, federal government, private market, national security, citizenry.
Citations:
Center for Research on Foundation Models (CRFM) and Stanford Institute for Human-Centered Artificial Intelligence (HAI). (2021). On the Opportunities and Risks of Foundation Models. Retrieved from https://crfm.stanford.edu/assets/report.pdf
Stanford Institute for Human-Centered Artificial Intelligence. (2023, April 12). The State of AI in 2023: 14 Charts. Retrieved from https://hai.stanford.edu/news/2023-state-ai-14-charts.
---
Send in a voice message: https://podcasters.spotify.com/pod/show/couragevconvention/message