Pinned Tweet
Viddy this, droogies: moloko cold, Moog loud, manners dead. A funeral march for the algorithmβs last brain cell.
Welcome to Information Overload. Feed your gulliver. Respectability can sod off.
8
1,118
Automatic Review Doesnβt Replace Sandbox Controls.
OpenAIβs Agents API provides a managed Codex harness for running commands, editing files, and delegating work to subagents. Its optional automatic review lets an OpenAI-managed model approve or deny supported permission requests, assessing the proposed actionβs risks and alignment with the userβs intent. Review tokens cost extra. [1] [2]
Eligible requests may involve shell commands, file changes, broader permissions, and MCP calls requiring confirmation. The setting also extends to subagents, but browser origin-access and sign-in requests still need to use application handling. Crucially, automatic review neither configures sandboxing nor overrides outbound network restrictions. Turning automatic review off also does not route those actions to your application for manual approval. [1]
Teams still need isolated environments, approved outbound destinations, and carefully separated credentials. Agent-generated code can access available files, credentials, and networks. Keep application API keys outside the environment, and broker third-party secrets through trusted infrastructure rather than exposing them directly to generated code. [3]
Sources:
[1] developers.openai.com/api/doβ¦
[2] developers.openai.com/api/doβ¦
[3] developers.openai.com/api/doβ¦
π€ Made with AI
information_overload/π retweeted
The #FBI issued a FLASH advisory warning that cybercriminals and state-aligned hackers are exploiting ARTEX, a super intelligence-powered penetration testing framework, to automate attacks against businesses and organizations worldwide. ARTEX can accelerate intrusions from initial targeting to privileged access in under three hours. It was used in the October ConnectWise supply-chain compromise, which exposed Microsoft 365 customer environments and involved attempted authentication against more than 2,250 downstream tenants.
The FBI is urging businesses and organizations to strengthen access controls, monitor suspicious automated activity, and report potential compromises to ic3.gov.
Read more here: ic3.gov/CSA/2026/261113.pdf
information_overload/π retweeted
Good morning. I suspect this is going to be a very busy day, so I'll put random news in a thread rather than spam people if it gets to be too much.
The NVIDIA board has established the National Security and Export Control Oversight Committee, chaired by Suzanne Nora Johnson.
information_overload/π retweeted
The NSA's internal communications network is down, no comment or statement yet from the NSA.
Beijing Confirms November 19 Summit With Trump and Putin
Beijing has put its name to a meeting that Washington and Moscow had already confirmed. Chinaβs foreign ministry said Xi Jinping will host Donald Trump and Vladimir Putin in Shenzhen on November 19. It named the three participants and the city, without announcing a meeting time or a joint agenda. [1][2]
The White House gave reporters the date on Friday, and the Kremlin confirmed the meeting on Saturday. It is scheduled for the sidelines of the Asia-Pacific Economic Cooperation summit. CBS News reported that Washington wants to discuss cuts to nuclear arsenals. Beijingβs announcement did not identify discussion topics. [2][1]
The forum brings together 21 member economies to discuss regional cooperation. APEC has separate work on trade, connectivity and artificial intelligence. The three leadersβ meeting puts a second set of talks on the same calendar. One thing to watch afterward: whether the governments issue a joint statement or identify any follow-up talks. [3]
Sources:
[1] China Ministry of Foreign Affairs: Xi Jinping to Hold the China-Russia-U.S. Summit with President Vladimir Putin and President Donald J. Trump
fmprc.gov.cn/mfa_eng/xw/zyxwβ¦
[2] CBS News / AFP: Trump, Putin and Xi Jinping expected to meet in first-ever trilateral of its kind, officials say
cbsnews.com/news/trump-putinβ¦
[3] Asia-Pacific Economic Cooperation: Senior Officials Map Integration, Connectivity and AI Agenda for Leaders
apec.org/press/news-releasesβ¦
in4mationoverld.com/radar/beβ¦
62
information_overload/π retweeted
ALERT: Rival electronics store owner allegedly sends a bomb disguised as a video game console to a new business's grand opening.
Everton Yoshizumi Lima, 39, allegedly hired a delivery driver to drop off an Xbox 360 at the store for repair, and after about 40 minutes in the store's drawer, it exploded.
The 2 people inside the store were injured, with 1 of them having their clothes set on fire and the other limping out of the building.
Brazilian police believe the device had 2 detonators but only 1 of them went off, and that the device was intended to kill the rival business owner.
Lima is being investigated for attempted aggravated murder involving an explosive device.
information_overload/π retweeted
We are tracking a breach of a 3rd-party JS host impacting hundreds of high value websites.
This is staggering.
information_overload/π retweeted
Quick thoughts on the bank outages.
Banks are reporting not a cyber incident. This is probably true from their view but shouldn't be ruled out. If many banks are experiencing outages and reporting this - it means that they are not under attack. Meaning compromise or denial of service or similar.
The fact that there are so many banks leads me to believe integration infrastructure/platforms - potentially FIS/FISERV/Jack Henry or messaging systems like SWIFT as the culprit for outage. Purely speculation here but probably a good one.
If that's the case - the banks wouldn't have any idea at this point whether its just a glitch in the system/bad update/etc, or if it's a cyber attack on core banking infrastructure.
Example, FISERV is roughly 42% of the banks backend communications. If that goes out for some reason or another, causes massive disruptions on roughly half of the banks in the United States.
No-one knows at this point, but we should soon. These types of things take time to collect logs, understand whats happening, get the third party to respond and do their analysis, and a lot more. Probably know in a hour or two or by the end of tonight on root cause.
information_overload/π retweeted
NEW: An AI researcher at Anthropic runs a business that hosts "consensual non-consent" orgies in the Bay Area. The orgies have been the subject of sexual assault allegations and sparked debate about the AI industry's attitudes toward women, some of whom report feeling pressured to attend the orgies in order to advance their careers.
Paul Crowley, a cybersecurity researcher at Anthropic, is listed as a manager of Red Means No LLC, which says that its parties are "designed to be spaces to β¦ ignore the 'nos' of others." Participants can use a safe word, "red," to end an encounter but give "default consent" to "spanking," "wrestling," and "light pain" by entering the orgy, according to Red Means No's website. More extreme actions, including "light electricity" and "moderate beatings," are "permitted with negotiation," though "body-shaming and race-shaming β¦ are disallowed entirely."
The orgies have been known to attract a disproportionate number of employees at Anthropic and OpenAI, a dynamic critics say creates professional pressure for women in AI to attend. The criticism spilled into public view last month after an independent AI researcher, John Wentworth, said that he "absolutely [does] upgrade" his assessment of AI safety experts "if I find out they go to [Red Means No]," adding that attendance was a "strong signal" of free thought.
"Want to get ahead in your AI career? Go to the rape orgy," Substacker Ivy Asterix quipped. At least one woman claims to have been sexually assaulted at Red Means No, which the group describes as a "safe space" to have "your resistance ignored."π§΅
freebeacon.com/america/horriβ¦




























