FTC Chair: Open AI models can “liberate startups”
The Commissioner of the FTC, Lina Khan, spoke at a Y Combinator event yesterday, supporting open AI models as a way to help level the playing field for “little tech” in the fast-moving industry.
Khan noted that at moments of technological disruption such as AI, “these are also the moments when incumbents can try to tighten their grasp, even if it means abusing their power because they have the most to lose.”
“Open weights models can reduce costs for developers so that they can focus their capital on products and services rather than expensive model training, and they can free up venture capitalists to pursue promising new applications of models rather than starting at square one with model development.”
-Lina Khan, FTC Chair
Khan made a point to clarify that the FTC supports “open weights models,” rather than AI models simply labeled as “open source,” which don’t make the model weights available. Model weights are basically the weighted contextual relationships between words. When weights are made available, others have more freedom to tailor “distilled” AI models from the larger model, and in Khan’s view “drive innovation, promote competition and consumer choice and lower costs and barriers to entry for startups like the ones that incubate here.”
This week’s release of Llama 3.1 from Meta is an example where the model weights were made available, in addition to the model itself.
“Open weights models can reduce costs for developers so that they can focus their capital on products and services rather than expensive model training, and they can free up venture capitalists to pursue promising new applications of models rather than starting at square one with model development.”
-Lina Khan, FTC Chair
Khan made a point to clarify that the FTC supports “open weights models,” rather than AI models simply labeled as “open source,” which don’t make the model weights available. Model weights are basically the weighted contextual relationships between words. When weights are made available, others have more freedom to tailor “distilled” AI models from the larger model, and in Khan’s view “drive innovation, promote competition and consumer choice and lower costs and barriers to entry for startups like the ones that incubate here.”
This week’s release of Llama 3.1 from Meta is an example where the model weights were made available, in addition to the model itself.
