Technology AI, Science and New Things
Named Grok, the chatbot reportedly comes with an attitude – and is a reflection of the billionaire’s wavering commitment to AI safety
No wonder Elon Musk was happy to make time in his schedule last week for Rishi Sunak. The PM wanted to chat about AI and Musk, it so happened, had an AI product ready to go. It’s called Grok. It’s being tested by select US users as a prelude to wider release, and he says it’s better than ChatGPT because it feeds off dynamic X (ie ex-Twitter) data as opposed to static data sets that cut off before the present and are therefore historic (and not in a good way).
Two points to note: Musk says Grok will channel the Hitchhiker’s Guide to the Galaxy. It’ll have attitude; a rebellious streak. It will play along with questions about how to make cocaine before advising against.
But secondly, earlier this year Musk signed a letter calling for a pause in AI development to allow for the development of shared safety protocols, only to post a message on his X site yesterday saying he signed the letter “knowing it was futile”. He “just wanted to be on the record as recommending a pause”. This is tech cake-ism distilled to its essence.