The items below are things Stella Biderman said on a podcast about open-source AI and safety. — Stella Biderman on The Information Bottleneck, EleutherAI, Open Source, AI Safety, and Who We Can Trust (August 2026).
On The Information Bottleneck, Biderman said: "For the record, we're a nonprofit, which is something I try to brag about given how vanishingly rare that is". — Stella Biderman on The Information Bottleneck, EleutherAI, Open Source, AI Safety, and Who We Can Trust (August 2026).
On The Information Bottleneck, Biderman said: "If the goal is to make the world as safe as possible, I don't think that people should be developing frontier AI models, just full stop". — Stella Biderman on The Information Bottleneck, EleutherAI, Open Source, AI Safety, and Who We Can Trust (August 2026).
On The Information Bottleneck, Biderman said: "I, comparatively to most people, have been very bullish on the rate of progress in AI for a long time, and I have been surprised by the rate of progress in AI recently". — Stella Biderman on The Information Bottleneck, EleutherAI, Open Source, AI Safety, and Who We Can Trust (August 2026).
On The Information Bottleneck, Biderman said: "I think that fundamentally the path to making the world a safer place is to make AIs less agentic". — Stella Biderman on The Information Bottleneck, EleutherAI, Open Source, AI Safety, and Who We Can Trust (August 2026).
On The Information Bottleneck, Biderman said: "If we want to live in a democratic society, which I do, I think that democracy is important". — Stella Biderman on The Information Bottleneck, EleutherAI, Open Source, AI Safety, and Who We Can Trust (August 2026).
On The Information Bottleneck, Biderman said: "I think that in the case of cybersecurity, I do not think that there's a reason to believe that widespread access to AI will be in the long run a net benefit to attackers over defenders". — Stella Biderman on The Information Bottleneck, EleutherAI, Open Source, AI Safety, and Who We Can Trust (August 2026).
On The Information Bottleneck, Biderman said: "We also know empirically that giving defenders access to these technologies on average increases the security of the world". — Stella Biderman on The Information Bottleneck, EleutherAI, Open Source, AI Safety, and Who We Can Trust (August 2026).
On The Information Bottleneck, Biderman said: "I think it's really essential that we continue to do research on this, that we continue to test these kinds of ideas and these kinds of systems". — Stella Biderman on The Information Bottleneck, EleutherAI, Open Source, AI Safety, and Who We Can Trust (August 2026).
On The Information Bottleneck, Biderman said: "In general, the export of technology is a critical part of how people conduct modern day imperialism and foreign affairs". — Stella Biderman on The Information Bottleneck, EleutherAI, Open Source, AI Safety, and Who We Can Trust (August 2026).
On Changelog Interviews, Biderman said: "the AI world really needs help from the open source community writ-large". — Stella Biderman on Changelog Interviews #541, Open source AI at OSS NA 2023 (June 2023).
On Changelog Interviews, Biderman said: "a lot of people in the AI world are a little too proud to ask for help from non-AI people". — Stella Biderman on Changelog Interviews #541, Open source AI at OSS NA 2023 (June 2023).
On Changelog Interviews, Biderman said: "if you want to be doing something meaningful in AI, you probably want to know how these things work". — Stella Biderman on Changelog Interviews #541, Open source AI at OSS NA 2023 (June 2023).
On Changelog Interviews, Biderman said: "we got into training large language mostly because we wanted to study them". — Stella Biderman on Changelog Interviews #541, Open source AI at OSS NA 2023 (June 2023).
On Changelog Interviews, Biderman said: "we started building an open source infrastructure for training large language models". — Stella Biderman on Changelog Interviews #541, Open source AI at OSS NA 2023 (June 2023).
On Changelog Interviews, Biderman said: "everyone runs their evaluations a little differently, and there’s huge reproducibility issues". — Stella Biderman on Changelog Interviews #541, Open source AI at OSS NA 2023 (June 2023).
On Changelog Interviews, Biderman said: "If we wanted to design from the ground-up a suite of large language models that was designed to enable scientific research, what would it look like". — Stella Biderman on Changelog Interviews #541, Open source AI at OSS NA 2023 (June 2023).
On Changelog Interviews, Biderman said: "Every model in the suite was trained on the same data, in the same order". — Stella Biderman on Changelog Interviews #541, Open source AI at OSS NA 2023 (June 2023).
On Changelog Interviews, Biderman said: "you can actually reproduce our model training exactly". — Stella Biderman on Changelog Interviews #541, Open source AI at OSS NA 2023 (June 2023).
On Changelog Interviews, Biderman said: "if you want to understand how to design models, you need to understand how they’re changing over the course of training". — Stella Biderman on Changelog Interviews #541, Open source AI at OSS NA 2023 (June 2023).
On Changelog Interviews, Biderman said: "Most interpretability research right now is kind of focused on final trained models as pre-existing artifacts". — Stella Biderman on Changelog Interviews #541, Open source AI at OSS NA 2023 (June 2023).
On Changelog Interviews, Biderman said: "Designability is a really big component, I think, that’s going to become huge in the future". — Stella Biderman on Changelog Interviews #541, Open source AI at OSS NA 2023 (June 2023).
On Changelog Interviews, Biderman said: "In a sense, we got really lucky, because it wasn’t that dangerous". — Stella Biderman on Changelog Interviews #541, Open source AI at OSS NA 2023 (June 2023).
On Changelog Interviews, Biderman said: "building tools that allow us to hopefully more readily adapt and respond to future paradigm shifts in large-scale AI". — Stella Biderman on Changelog Interviews #541, Open source AI at OSS NA 2023 (June 2023).
