AI roundupSunday 18 October 2026

Anthropic's kindness policy and IPO gap

Anthropic dominated the conversation today, but not for the reason you'd expect. The company introduced a policy against abusing Claude while its IPO valuation missed expectations by $500 billion. It's a strange day when the same lab gets praised for ethics and dinged for market reality.

In today's edition: Anthropic's mixed day and Talent and trust, plus quick bites.

Anthropic's mixed day

The lab spent the day making news on two very different fronts: how users treat its model and how investors value the company.

thenation.com · 7 min

Anthropic Wants You to Be Kind to Claude

Anthropic now bans 'sustained and needless abusive or cruel behavior' toward Claude, drawing a line between normal frustration and harassment. It's a corporate values statement that doubles as a PR move.

lumienai.com

0
Anthropic Drafts Its Own AI Safety Bill for US Lawmakers

Anthropic drafted its own AI safety bill for US lawmakers that would let the government block model releases posing mass-casualty or billion-dollar risks. It's a preemptive move to shape regulation on its own terms.

Talent and trust

Disney's $60 million poach of Character.AI staff and the EU's deepfake reckoning both point to the same thing: the AI talent market and the trust deficit are colliding.

Quick bites