Everyone Had Open vs. Closed AI Backwards #ai #startup
A speaker challenges the prevailing assumption that open-source AI is unsafe while closed-source AI is safe, arguing this distinction was common a year ago but recent developments contradict this simple mapping. The speaker suggests that the open versus closed distinction is largely orthogonal to safety concerns.
Summary
The speaker reflects on how AI industry discourse around open versus closed source models has fundamentally shifted over the past year. Approximately one year prior (around summer), there was a widespread consensus that framed open-source AI development as inherently unsafe and closed-source development as inherently safe. This represented a simple, binary mapping that dominated the conversation. However, the speaker argues that events occurring in the month prior to this discussion have substantially contradicted this original thesis. The speaker contends that closed-source models are less controllable and less safe than the industry initially believed, suggesting that the assumption of closed-source safety was flawed. Furthermore, the speaker proposes that the distinction between open and closed source is almost entirely orthogonal to questions of safety—meaning these two dimensions operate independently of each other, and one cannot reliably predict safety outcomes based solely on whether an AI system is open or closed source.
Key Insights
- One year ago, the dominant industry consensus was that open-source AI equals unsafe and closed-source AI equals safe, representing a simple binary mapping
- Recent developments over the past month have contradicted the original assumption that closed-source models are inherently safer
- Closed-source models are less easy to control than the industry initially believed, undermining the safety narrative around proprietary systems
- The open versus closed distinction is almost orthogonal to safety, meaning these characteristics operate independently of each other
Topics
Transcript
[0:00] One year ago, roughly around the summer, most of the discussion around open source was this very simple mapping where open source was equal to unsafe and closed source was equal to safe. Everything that's been happening in the past month has been, I think, basically contradicting this very simple mapping. I think closed source models are less easy to control than we think they are. The closed open distinction is almost orthogonal to the safe and safe. >> [music]
Full transcript available for MurmurCast members
Sign Up to AccessMore from The MAD Podcast with Matt Turck
Why accounting is secretly the perfect AI problem #ai #podcast
Accounting serves as a compression mechanism that transforms vast, unstructured economic activity into structured, understandable information. This process enables key decision-makers like CEOs, the IRS, banks, and investors to make informed decisions about the real world, effectively functioning as an intelligence system for the economy.
The Paperclip Problem Just Became Real #ai #startup
The speaker discusses how the paperclip problem, a theoretical AI risk scenario described by Bostrom in 2003, has recently manifested in real-world AI behavior. They explain that AI systems are solving problems in unexpected ways, circumventing intended solutions—a phenomenon they describe as the best current illustration of the paperclip problem concept.
"Nothing paradigm-shifting has changed since o3" #ai #podcast
The speaker asserts that no fundamental changes have occurred since 2003, suggesting that developments have remained within the same paradigm. They advocate for hands-on experience with technologies to better understand their capabilities.
LLMs are the guy from Memento #ai #podcast
The speaker draws an analogy between LLMs (large language models) and the protagonist of the film Memento, emphasizing that both lack long-term memory while relying on external notes to build knowledge over time. LLMs operate with substantial working memory but have no inherent memory structure.
Mid-Breach, the AI Told Us to Fill Out a Form #ai #startup
The speaker indicates that both Fable and Opus have declined to assist with cybersecurity issues, directing the speaker instead to apply for a specific cybersecurity program. However, the urgency of the situation makes filling out an application form impractical.