TextJul 31
OpenAI disrupted a ChatGPT-assisted scam network
Users, businesses, and safety teams got a concrete example of how fraud operations can use ChatGPT for support. On July 31, OpenAI said it disrupted a Cambodia-based network involved in investment, romance, gambling, and law-enforcement impersonation scams, banned coordinated accounts, and shared threat signals with partners and authorities.
Not scored - This documents misuse and enforcement activity, without a new capability or access change.
Announced or reported, but nothing you can use or verify. Some of these will become real. Some won’t.
AgentsAug 1DeepSeek V4-Flash reportedly adds agent APIs
Waiting on: Primary verification - A single newsletter account is the supplied source; an official DeepSeek release page or matching reports from multiple authoritative outlets would verify the claimed public access.
RobotsJul 31KUKA reportedly deploys AI management at Ohio factory
Waiting on: Primary verification - One trade-publication report describes the deployment; a KUKA announcement or matching reports from multiple authoritative outlets would confirm the commercial operation.
ScienceAug 1OpenAI reports Astra results on longstanding math problems
Waiting on: Peer review or independent replication - The supplied material reports internal-model results without peer review or independent replication; a peer-reviewed publication or independent replication would meet the science bar.
