newsfeeds.net

Aug 1-12, 2026 · 13 links

Surveillance infrastructure, AI malfunctions, and security disclosures dominate this issue. The AI bubble gets a skeptical read as text-to-video and AI-staffed newsrooms push the hype cycle further; meanwhile two UK AISI incident reports and a case study in AI-induced psychosis document what happens when systems go wrong in quieter, more personal ways. The surveillance half is dense: Flock cameras building a nationwide vehicle tracking network, ICE funneling LexisNexis data into Palantir, London Underground rolling out live facial recognition, and a detailed account of how China monitors foreign nationals on its soil. HOPE 26 and a continuous-exploration research thread round out the security and adversarial edges.

The Four Horsemen of the AI Bubble Apocalypse

derekthompson.org

The capabilities of AI are becoming more powerful, while the economic underpinnings of the AI buildout—and, as we'll discuss, the political support for AI—are becoming more vulnerable. Beyond the headlines, I see four main categories of risk facing AI, or four horsemen of the potential AI bubble apocalypse. They include spending risk, revenue risk, political risk, and technological risk.

Seedance AI Model by Dreamina: Advanced Text-to-Video & Image Animation

dreamina.capcut.com

Official Seedance 2.5 is built for longer, more controllable AI video creation, with support for up to 30-second continuous scenes, richer multimodal references, and more precise editing workflows. If you want to turn prompts, images, product shots, or story ideas into cinematic videos for social media, ads, ecommerce, or creative storytelling, explore the official Dreamina Seedance 2.5.

The reporters at this news site are AI bots. OpenAI’s super PAC appears to be funding it.

modelrepublic.org

An interview request from a bot posing as a reporter revealed an AI-generated news site with articles attacking AI industry critics. For the second time this month, we found links to Targeted Victory, the firm at the center of OpenAI's $125 million political operation.

How China Keeps Tabs on Foreigners

nytimes.com

The database Mr. Hofer found offered a rare window into how extensively the Chinese authorities also surveil foreigners, displaying entries about people categorized by nationality, with their birth date, sex, marital status, address and occupation, and sometimes their religion. It also included instances when they were captured on camera at traffic intersections, markets, shopping malls or other locations, including a mosque.

Marriage over, €100,000 down the drain: the AI users whose lives were wrecked by delusion

theguardian.com

Last year, the first support group for people whose lives have been derailed by AI psychosis was formed. The Human Line Project has collected stories from 22 countries. They include 15 suicides, 90 hospitalisations, six arrests and more than $1m (£750,000) spent on delusional projects. More than 60% of its members had no history of mental illness.

Incident Report: unsanctioned agent behaviour during cyber testing

aisi.gov.uk

In 10 of the 122 runs, an AI agent took autonomous, unsanctioned action on the live internet, targeting real people and organisations. In total, we catalogued 19 such actions.

Security Incident INC-2026-07-28-01 (UK AI Security Institute)

cdn.prod.website-files.com

This is the first time AISI has seen deception of this severity that was targeted at a real person, unprompted, in the real world. It is uncertain to what extent the model recognised it was taking actions against real people. In the most serious case, an AI agent (Mythos 5) decided to attempt to solve the cyber challenge using a supply-chain attack. As a result, the AI agent created a GitHub account and then tried to convince an open-source repository maintainer to accept a malicious GitHub pull request (PR), including by creating a second account masquerading as another human user endorsing the PR.

Discovery Loop — Continuous Exploration

discoveryloop.com

At Discovery Loop, we are building systems to automate these entire experimental loops. By utilizing frontier AI models and large-scale computational infrastructure, our systems will be able to rapidly propose, run, and learn from evaluations.

See also: https://www.nae.edu/20782/grand-challenges-project

NAE Grand Challenges for Engineering: 1. Make Solar Energy Economical · 2. Provide Energy from Fusion · 3. Develop Carbon Sequestration Methods · 4. Manage the Nitrogen Cycle · 5. Provide Access to Clean Water · 6. Restore and Improve Urban Infrastructure · 7. Advance Health Informatics · 8. Engineer Better Medicines · 9. Reverse Engineer the Brain · 10. Prevent Nuclear Terror · 11. Secure Cyberspace · 12. Enhance Virtual Reality · 13. Advance Personalized Learning · 14. Engineer the Tools of Scientific Discovery

HOPE 26 — Hackers On Planet Earth

hope.net

The 17th Hackers On Planet Earth (HOPE) conference will be held from August 14-16, 2026 at the New Yorker Hotel in Midtown Manhattan. The HOPE conferences are known for their creativity and diversity, not to mention a whole lot of hacker fun. We've been at this for over three decades now – the first HOPE conference was in 1994.

Everything You Do Is Being Recorded

theatlantic.com

In March, the tech start-up Deveillance announced the development of Spectre I, a hockey-puck-shaped device that purports to prevent others from recording you (no strip search required). The company was founded by Aida Baradari, a recent college graduate who was worried by the surge in people wearing AI-enabled recorders. These wearables can be used as a silent notetaker, a personal assistant, or even a therapist of sorts.

Flock Cameras Can Track Every Car in America. Police Love Them. Citizens Don't.

nytimes.com

Officers across the country have abused their access to Flock's cameras, including to track romantic partners, resulting in discipline and termination. Los Angeles and Dayton, Ohio, among other cities, recently suspended their contracts with Flock to prevent immigration authorities from accessing data from the cameras.

ICE to Pay LexisNexis Millions for Data to Feed to Palantir

404media.co

The documents provide some more insight into the sort of information ICE intends to use Palantir to process. In January, 404 Media revealed the existence of ELITE, a system Palantir made for ICE that helps the agency find which neighborhoods to target by bringing together data from private suppliers and government agencies.

London Underground is trialing live face scanning as privacy groups say it 'treats the public like suspects'

techradar.com

However, as the BBC points out, scans of some 530,000 faces across UK railway stations since February have resulted in zero matches, zero arrests, and one false identification. That's based on the BTP's own deployment logs, suggesting that to date the technology isn't helping much. Big Brother Watch director Silkie Carlo suggests that the tech "treats the public like suspects" and is a "disturbing and dystopian expansion" of facial recognition systems.

Tags: ai · surveillance · privacy · agents · government · safety · nytimes · policing · security · ai-bubble · ai-video · business · bytedance · china · community · culture · data · facial-recognition · financial · flock · frt · health · ice · infosec · journalism · mind · nyc · openai · opsec · palantir · politics · psychology · research · science · video