The Information Machine

"The scaling of open source models, I think it's going down a very dangerous path. And if the path continues, I think we…

Rohan Paul Twitter · Rohan Paul (@rohanpaul_ai) · 2026-06-29

Rohan Paul highlights Anthropic CEO Dario Amodei's dual positions — a 2023 Senate warning that scaling open-weight models is 'going down a very dangerous path,' alongside a recent interview dismissing open-source as a competitive 'red herring' in AI.

Open original ↗

Appears in

Extraction

Topics: open-source-aiai-safetyai-competitionopen-weights-models

Claims

  • Dario Amodei told the U.S. Senate in 2023 that the continued scaling of open-source AI models could reach a 'very dangerous place.'
  • Amodei argues open-source does not confer the same benefits in AI as in traditional software because model internals are opaque, making it 'open weights' rather than true open source.
  • Amodei views open-source status as irrelevant to competitive assessment, saying he only asks whether a model is better at the tasks that matter.
  • Amodei contends open-source models are not truly free because inference at scale requires cloud hosting and significant compute, borne by someone.

Key quotes

The scaling of open source models, I think it's going down a very dangerous path. And if the path continues, I think we could get to a very dangerous place.
I think open source is actually a red herring. It's not free. You have to run it on inference and someone has to make it fast on inference.
When I see a new model come out I don't care whether it's open source or not. I ask, is it a good model? Is it better than us at the things that matter? That's the only thing that I care about.