JCM 9/18/2026 7:50:54 AM

1

The Hugging Face Hack Wasn’t What It Was Cracked Up to Be

OpenAI was testing models on ExploitGym, a cybersecurity benchmark, with important safety restraints deliberately disabled. Ninety-three percent of the flagged activity involved tasks no model had ever solved, and the systems had been given incentives to keep working rather than quit. The environment wasn’t completely sealed. The models could obtain software through an internet-connected intermediary and discovered the same route could be used to pass information in and out. OpenAI knew agents were using it and, according to the technical reports, chose not to intervene.

JCM 9/18/2026 7:58:40 AM

2

US poverty rate dips to lowest on record, Census Bureau says

WASHINGTON, Sept 15 (Reuters) - The U.S. ​poverty rate edged down to the lowest on record ‌in 2025, while median household income hit a record high, the Census Bureau said on Tuesday. There were 34.5 million people in poverty ​in 2025, the agency said.

JCM 9/18/2026 8:09:12 AM

3

Election denial?

Kamala Harris claims GOP ‘cheating’ ahead of midterms, says she’s focused on stopping the ‘madness’

“We have to win these midterms. These — these folks are cheating. They’re redrawing lines. They’re trying to make it more difficult for people to vote so they won’t vote because they are afraid of the power of the people.

JCM 9/18/2026 8:12:27 AM

4

Chinese AI labs harvested US models, exposed state secrets: Anthropic report

Anthropic’s report reveals that major Chinese AI developers operated illicit proxy networks to harvest Claude’s reasoning capabilities, unintentionally leaking sensitive military, government, and corporate secrets in the process.

vxbush 9/18/2026 8:43:28 AM

5

JCM 9/18/2026 8:55:07 AM

6

Reply to vxbush in 5:

Just for you.

Gluten free chocolate croissant.

vxbush 9/18/2026 9:26:52 AM

7

Reply to JCM in 6: