“Byrne Hobart believes that simply starting another AI safety-focused lab may not be an effective strategy, as it would likely struggle to compete with the resources of entities like OpenAI or Google.”