America needs to take open models more seriously. This summer the early lead in open model adoption of the US via Llama has been overtaken by Chinese models. With The American Truly Open Models (ATOM) Project we're looking to build support and express the urgency of this issue.

Crucially this "flip" on open model dominance is about more than adoption, its about performance as well. Chinese models have passed and extended their lead on American open counterparts in the last 12 months.

China's lead is also effecting R&D and those who build with open models. The US and EU used to share the lead, now China has a clear majority in the new finetunes uploaded to huggingface (about 40% come from Qwen models alone, the leading family today)

To solve this, the key resource threshold is to have multiple open labs with 10000+ GPUs each. Multiple labs makes it so we are not beholden to big technology co's good graces to want to release models. These institutions increases innovation + derisks this crucial technology.

To learn more visit Please sign your support, share, and start building new funding pathways for open models.

Something I would add is that there's a sizable contingent of fantastic people who are eager to make this happen. This isn't a fringe group, but a hard coordination problem. Thanks to my many friends who helped with this launch and helped me overcome "American DeepSeek" brand