Social Comms

We have more to say than we have places to say it. NOPE runs open evaluations of raw and deployed AI, maintains an incident record and a regulation tracker, and publishes original research on relational safety. Almost none of it reaches the people who would care, because nobody owns our channels.

The job is to change that. You'd take what we produce (benchmarks, incident analysis, taxonomy work, the occasional strong opinion) and turn it into a consistent, human presence on the channels that matter: deciding what gets posted, writing it, engaging with the researchers, journalists, builders and policymakers already wrestling with these questions, and telling us honestly what's landing and what isn't.

We're making AI safer in a way that is less about terminators and bio-weaponry and more about the human being at the centre of these relationships. The labs at the frontier of brute intelligence are not at the frontier of emotional, psychological and social intelligence, and we're holding them to account without shouting from the sidelines. We build open-source tooling and models to fill the gaps we see, we run evaluations of raw and deployed AI, and we published the first framework for relational safety. It goes beyond 'do no harm': AI that advises you while supporting your autonomy, a companion without endless praise, honest about its abilities and identity, and supporting your wellbeing in the long term, not just turn by turn.

The measure of the work is whether the right people are actually engaging: researchers, builders, journalists, policymakers, and the ordinary users who sense something is off. No growth-hacking, no engagement bait. We would rather stay small than sound like a brand. The work is async, any timezone, and you'd have direct access to the founders and to everything we publish before it's public.