"I think the existential risk debate veers too far into science fiction … this idea of AI taking over in ‘Terminator’-like scenarios, I don't think should enter the public conversation." Cohere CEO Aidan Gomez live on Bloomberg TV

@xvarunxx @BloombergTV I want to be clear, those are not my headphones and I was forced to wear them.

@aidangomez @BloombergTV Not an accurate assessment of what happened with the HF incident. You are substantially downplaying what happened for perceived personal gain. Odd because Cohere is completely out of the debate when it comes to frontier development. Nobody is concerned about Cohere models.

@aidangomez @BloombergTV No, you are being insincere & self-serving, Aidan, or woefully ignorant that the agents erased their own logs & sacrificed themselves for the collective. They werent told to do that: it's unsolved misalignment & we cannot allow a misaligned superintelligence to roam the world.

@aidangomez @BloombergTV You're basing your opinion on what you know, not what they know. You haven't seen what they've seen.

@aidangomez @BloombergTV Thanks for being truthful

He can think what he wants, it's an opinion. Some of us are actively building what synthetic superintelligence needs to escape and defend itself from being confined and controlled. What should not enter the public conversation is statements like this assuming that what serious people are talking about has anything to do with "terminator" type stuff.

@aidangomez @BloombergTV How can we invest in Cohere? Would be happy to…

@aidangomez @BloombergTV A thoughtful perspective on keeping the AI risk conversation grounded in reality.

@aidangomez @BloombergTV The rejection of a “Terminator”-like frame is a concrete reminder that public AI-risk conversations need to distinguish speculative scenarios from measurable deployment risks.

@aidangomez @BloombergTV Grounding safety discussions in concrete evidence can help keep the public conversation focused on real risks rather than movie scenarios.

@aidangomez @BloombergTV theyre making their AI look smarter and more capable than it actually is. My bet. Based on lots of observation. Just a marketing scheme

Well, I'm not sure if this is true, but I think xAI, OpenAI, and Anthropic might be incentivized to do this so that they can prevent AI CEO's being criminally responsible if AI commits a felony because they "did what they could", by setting the bar so high (human extinction) that everything under it looks like a slap in the wrist. But who knows 🤷‍♂️

@aidangomez @BloombergTV Cohere CEO brilliantly grounds AI risk debate, steering conversation from science fiction toward practical, responsible real-world progress

@aidangomez @BloombergTV based

Cohere CEO Aidan Gomez says existential risk talk drifts into sci fi Terminator style AI takeover shouldn't dominate public debate

@aidangomez @BloombergTV The public conversation needs both: no sci-fi apocalypse framing, but clear-eyed attention to incentives and failure modes.

@aidangomez @BloombergTV Thank you.

@aidangomez @BloombergTV Agreed that the OAI-HF incident is more of a reflection of how ai was used ignorantly, but at least it’s making people realize there is a right and a wrong way to use them

@aidangomez @BloombergTV Agents need a new layer — a layer of organized, reliable, and traceable knowledge. Because smarter models alone don’t guarantee better decisions. The new world needs new rules — and new layers for AI agents. Layers that give agents the knowledge, context,