Project Magi, X.AI, and GPT-4 improvements mean AI’s frenetic pace not slowing

The trend: The generative AI arms race isn’t slowing down.

The fear of missing out: Companies continue ratcheting up commercial generative AI efforts despite a six-month moratorium letter that Musk signed. OpenAI CEO Sam Altman didn’t sign the letter, criticizing it over a lack of “technical nuance,” per The Register.

  • OpenAI says it hasn’t begun GPT-5 training, but the difference between not training GPT-5 and enhancing GPT-4 could amount to a semantic sleight of hand.
  • Musk’s X.AI ambitions beg the question: How could his chatbot compete with ChatGPT unless it's based on a more powerful model?
  • Google’s precautionary instincts over the technology are getting bulldozed by market forces.

What’s next? Google’s internal testing that shows generative AI’s startling capacity for spontaneous learning illustrates the technology’s status as a wild animal yet to be tamed by regulation.

  • We’ll likely see more outcry among those in the AI field who say we’re getting closer to the precipice of catastrophe.
  • Nvidia’s H100 release and new AI training techniques that might require less compute power mean that the cloud server crunch won’t throttle the pace of development for long.
  • As the EU’s AI Act gets closer to passage, we’ll likely see more lobbying efforts for lax restrictions from some tech companies, while others might quietly want stricter regulation to provide justification for caution.
  • Because no one understands how advanced AI models work, they might be difficult to effectively regulate, hence calls to “shut it all down.”