Startup founders ask the U.S. government to keep Chinese open‑weight AI online
Several tech founders have sent a public letter to the U.S. administration.
They warn that shutting down Chinese open‑weight AI models could hurt global innovation.
The founders say the models are used by researchers worldwide for testing new ideas.
Closing them may push work into closed systems, slowing progress for everyone.
The letter also points out that open‑weight AI helps improve safety tools.
Developers can inspect the code, find bugs, and add safeguards.
A simple Python example shows how an open model can be loaded and examined:
from transformers import AutoModelForCausalLM, AutoTokenizer
model_name = openai/chinese-open-weight
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name)
inputs = tokenizer(Hello, world!, return_tensors=pt)
outputs = model.generate(**inputs)
print(tokenizer.decode(outputs[0]))
The founders ask the government to consider the broader impact before any ban.
They urge a balanced approach that protects security while keeping research open.
Top comments (0)