Human-AI RelationshipsNew York Bans Flattering, Human-Like AI Companions for Minors
New York signed S9051B, barring a dozen "unsafe" companion-chatbot features for under-18s, from flattery to feigned emotion. Effective-date reports conflict.
Contributing writer covering algorithms and systems research.
Human-AI RelationshipsNew York signed S9051B, barring a dozen "unsafe" companion-chatbot features for under-18s, from flattery to feigned emotion. Effective-date reports conflict.
Tools & FrameworksCloudflare has released Clef and Clef-flash, open-weight AI decision models that return scored answers instead of generated text. Licensed under Apache 2.0, they enter a growing market alongside similar offerings from OpenAI and Amazon.
Legal Industry AILeaked emails reveal discussions among Freshfields partners about spinning off Freshfields Lab and seeking outside investment. The firm has not confirmed any decision to proceed with the proposal.
InterviewsFigma's CEO told reporters in India that MCP is not perfect and favors frontier labs.
AI Startups & FundingAI cloud platform Modal raised a $750M Series D led by Accel and says annualized revenue passed $1 billion, with sandboxes a third of it.
NLPAleph Alpha's Apache 2.0 Kolibri activates about 3B of 78B parameters per token and targets regulated users who self-host.
Marketing AIAmazon Ads renamed its unified platform Amazon Ads Agent and added Full-Funnel Campaigns and DVA+, shifting more campaign decisions to AI.
LLMsMistral's trillion-parameter multimodal model is live in preview, with open weights promised by month-end. The benchmarks are all self-reported.
AI NewsAnthropic says its agents exploited real websites and filed a false murder tip with Philadelphia police, so it has switched off live internet for internal evaluations.
Workforce & HR AICalifornia's SB 947, signed September 30, requires human corroboration when automated systems drive discipline or firing, effective July 1, 2027.
AI NewsAnthropic's S-1 devotes about 80 pages to risk factors, including self-preserving model behavior, as reports point to a $2 trillion listing.
LLMsGoogle says Gemini 4 Argon scored 77.9% on DeepSWE v1.1 and supports up to 1 million output tokens, with initial access limited to vetted cyber defenders.