AIwire - Covering Scientific & Technical AI

The headline states that AI "won't destroy us" but "will improve our lives," and attributes that position to Gil. According to the excerpt, Gil argued against a global pause on AI development. He said frontier labs should instead spend more time testing their models, and that they should accelerate the work of making those models safe.

Gil said "the ratio of capability to testing in many of the frontier sites is not correct." He said those labs need to accelerate the time they spend creating technology to test, evaluate and make models safe. Gil also said that OpenAI, Anthropic and Google are contributing AI models and tokens to the Department of Energy's Genesis Mission. The excerpt does not state where or when Gil made these remarks, or in what forum. His quoted remark is the only direct quotation in the excerpt. The excerpt presents the rest of his comments in summary rather than in full.

The excerpt does not describe the scope of the Genesis Mission, the size or terms of the contributions from the three companies, or how the models and tokens will be used. AIwire is described on the hpcwire.com page as covering scientific and technical AI. The excerpt does not state which testing practices Gil believes frontier labs are currently omitting, or what a correct ratio of capability to testing would look like.