6 Comments
User's avatar
Thomas Dai's avatar

Banning will never solve the divide, it will make Anthropic and OpenAI more hated, it will widen the gap of already fragmented silicon valley.

When the cost difference is so obvious, people will always find a way to slip through the barrier, no matter how tough the policy is. It will just promote more usage in a hidden way.

The only right way forward is to embrace and have better business model

Michael Spencer's avatar

What are the chances China would ban the US from using their frontier open source models? I think that might be more likely the way things are going.

Tony Peng's avatar

So far I don't see that coming in the near term, China's policy is more retaliatory than proactive, and open models haven't reached the choke point level.

David G Brown's avatar

You make a number of good points, but ultimately aren’t these models just the tip of the spear for future Digital Belt and Road policies and much more geopolitically motivated. Under the guise of low cost and competitive open weight/open source, these products are right out of the standard export Chinese playbook to promote adoption at subsidized prices.

Tony Peng's avatar

I won't deny the possibility, but so far open-weight AI models have succeeded bc of competition and architectural innovation, not subsidies or national policy. The best move for American companies is to invest more in open models and compete.