Bloomberg reported that Harvey’s margins swung sharply as usage of its AI agents rose, and that the company and others have since shifted towards open-weight or in-house models to cut costs and dependence on frontier labs.
Read briefBloomberg : Sensor Tower: Muse was downloaded 902K+ times in the six days after its launch on Sept. 8, vs. Meta AI's 773K in the same post-launch period; META jumps 12%+ — Meta Platforms Inc.'s new artificial intelligence agent, Muse, has quickly risen to the top of mobile app…
Read sourceOpenAI said it had trained a new internal model, claimed major progress on open problems in mathematics, and said an independent advisory group would help guide how results are reviewed and shared.
Read briefThe Information : Source: before the Hugging Face incident, OpenAI was negotiating a legally binding deal with Anthropic for the companies to stress-test each other's models — OpenAI is rethinking a range of safety strategies as it responds to fears from employees and others…
Read sourceOpenAI said coding agents were already changing research work inside the company while it argued that any move towards fuller RSI had to preserve human control and stronger safeguards.
Read briefThe UN science panel on AI said its first report on the subject did not assure human control over AI agents, citing an OpenAI-linked incident and wider safety concerns.
Read briefMaria Curi / Axios : Ahead of Sam Altman's UN address, OpenAI urges the US to lead an effort to develop global safety and security standards for building frontier systems — OpenAI on Monday released international AI safety standards as world leaders, namely the U.S. and China…
Read sourceSoftBank planned a bond sale worth more than $11 billion, split between dollar and euro tranches, to make another payment on its OpenAI holding, according to the Financial Times.
Read briefThe group said it would advise AI companies on mathematical research, publish recommendations openly, and help with OpenAI’s release of results that it said came from its internal model.
Read briefThe article argued that international technical standards should guide frontier AI development, especially automated AI research and recursive self-improvement, and said the United States should lead that effort with other countries.
Read briefExplore new OpenAI Academy learning paths for employees, developers, leaders, educators, and students to build and demonstrate practical AI skills.
Read sourceResearchers found two ways out of OpenAI Codex’s sandbox, including one that could trigger host commands from read-only mode, and OpenAI patched both within eight days of being told.
Read briefCISA said three Linux kernel vulnerabilities were being exploited, set a same-day deadline for federal mitigation, and disclosed no details about the incidents or the threat actors involved.
Read briefResearchers published technical details and a proof-of-concept for Click2Shell, a WordPress Core CSRF chain that could lead to server-side PHP execution and was fixed in WordPress 7.1.1.
Read briefA joint advisory from agencies in Japan, the U.S., Australia and Germany said North Korean actors used Contagious Interview to compromise devices, steal wallet access and drain cryptocurrency.
Read briefShinyHunters said it had seized Cl0p’s leak site, demanded an extortion payment and linked the dispute to Oracle’s E-Business Suite campaign and earlier threats against its members.
Read briefSecuronix said a campaign it named TASK#STOMP used VBScript and PowerShell stages to persist on Windows hosts, collect sensitive data and maintain two command-and-control channels.
Read briefSpecops Software’s sponsored explainer said the FBI’s CJIS Security Policy v6.1 tightened encryption and vulnerability-scanning rules, while leaving password and MFA requirements largely unchanged and keeping local audit timing in place.
Read briefMicrosoft told Entra ID administrators to move users to phishing-resistant methods such as passkeys before SMS first-factor sign-in ends in February 2027, with QR codes and FIDO2 keys also listed.
Read briefBlackpoint researchers said a newly identified remote access trojan used ClickFix lures, hidden scripting stages and a Polygon smart contract to find active command-and-control servers, while the wider article also described other ClickFix campaigns.
Read briefSentinelOne said it found FLATROOF and ROOFDECK on an Apple Silicon MacBook used by a DevOps engineer at an India-based IT services provider, and linked the activity to Jade Sleet.
Read briefCheckmarx said an npm campaign centred on indexed-btree hid its loader in normal runtime code, let the package appear clean at install, and used extra packages to spread the operation.
Read brief