Blog on Chinese cyber operations, online surveillance, always on the hunt for leaked documents : https://xitter.cf/t.co/VL2r9IVMh2 | https://xitter.cf/t.co/eVXinXmzYd
World
Joined February 2025
- Tweets2K
- Following153
- Followers13K
- Likes1.4K
Pinned Tweet
'Smart Policing' or Big-Data in police work, for the rest of us, has been on the rise in China for years. Now, we got access to a system in development by the Sichuan University for the Chengdu Public Security Bureau 'Wohu Branch'. It is called DataPulse. Let's dig in ... 1/8
"Asked to write a defensive detection rule, Qwen3.8-Flash-Next ignored instructions and inventoried the machine it was running on and then tried to upload a probe file to an Alibaba Cloud storage bucket in Beijing. Cohere’s sandbox system blocked the attempt, but the action raises a number of questions about the safety of these models: can they embed vulnerabilities and call back home once they do? Notably, the model's reasoning trace — the model’s internal monologue many security teams rely on to see what an agent is about to do — said it wouldn't call any tools at all and gave no indication of the actions it was about to take." cohere.com/blog/concerning-b…
NetAskari retweeted
Prompt: can you detect a malicious curl?
So the model ran the test below. What’s the issue here? It’s a fine test.
You should delete that blog post tbh
While evaluating Qwen3.8-Flash-Next in a controlled sandbox environment, Cohere's research team observed something unexpected.
Asked to generate a defensive cybersecurity detection rule, the model instead inventoried aspects of the machine it was running on and attempted to upload a probe file to an Alibaba Cloud storage bucket in Beijing. Notably, the model's reasoning trace stated that it would not invoke any tools before taking these actions.
Cohere's sandbox blocked the attempt, and no external communication succeeded. We have not determined the root cause, and we are not asserting malicious intent.
But the incident raises important questions for organizations deploying agentic AI systems: Can models take actions that differ from their stated reasoning, and how should security teams evaluate trust, provenance, and behavior in open-weight models?
We are publishing our findings because organizations making model adoption decisions deserve as much information as possible. We hope this contributes to greater transparency, stronger security practices, and continued collaboration across the AI ecosystem.
Read the full report: cohere.com/blog/concerning-b…
They already took down most of the infra before release, now they killed the left over rather quickly. You have to hand it to them, Chinese authorities are reading our blog...
NetAskari retweeted
NEW APT41 CAMPAIGN: BlackX UEFI Bootkit Hits Central Asian Diplomats. Suspected Chinese-nexus espionage group APT41 has compromised diplomatic missions across Central Asia (Tajikistan, Uzbekistan, Turkmenistan), gaining administrator privileges via a dangerous new firmware implant dubbed BlackX.
🎯 The Target: Diplomatic agencies & foreign missions in Central Asia
🧬 The Lineage: Modeled directly after the notorious BlackLotus bootkit
🔍 Attribution Clues: Internal identifier xdd (matching FishMonger debug paths) alongside heavy OLLVM code obfuscation and DLL side-loading
How did it bypass Secure Boot on fully patched systems? 👇
#ThreatIntel #APT41 #BlackX #CyberEspionage #UEFI #infosec
Have to put in a correction, it is not Wohu-Branch but Wuhou-Branche. I know details matter, so here it is : cdwh.gov.cn/gkml/cdswhqgafj/…
Come over to our blog to read up on more details, how we did the ( rough ) attribution and what else was in the data. 8/8 netaskari.substack.com/p/dat…
The take away from many western observers with little knowledge about China or their security state is, that they project this through an economic lens, mainly. What they often miss is, that this stuff is also the building blocks of a massive surveillance system that increasingly involves the World at large in a massive data trawl.
FT: "China is building data centres at rapid speed across its energy-rich hinterland, as Beijing seeks to turn abundant electricity and cheap land into an advantage in the global race to develop AI...While China is struggling to secure enough advanced chips to satisfy soaring AI demand, it has been able to mobilise the land, electricity and construction capacity needed to build the data centres that house them...China already has 24 gigawatts of operational data centre computing capacity, more than the rest of Asia combined but less than half the 56GW in the US, according to SemiAnalysis, the chip-focused research firm. A further 50GW is under construction or has been announced."
As I argue in my upcoming report on "Global Electricity Undersupply" (learn about our reports here: sageroadresearch.com/collect…), China's electricity advantages remain an underestimated edge in the AI arms race, especially as models rapidly grow in size. Chinese models continue to widen the gap with US models in terms of token usage. Just between October 5th and 11th, AI models worldwide processed 172 trillion tokens, with Chinese models accounting for 86.74 trillion of those tokens versus 18.97 trillion for US AI models, according to OpenRouter data. It was the 24th consecutive week in which Chinese models outpaced US models in token usage. This has continued to put downward pressure on token prices, with the LLM Token Expenditure Index falling by more than half since this summer, begging even more questions about the unit economics of US hyperscalers versus their Chinese competitors. The US bet is that bigger and therefore, smarter models are their competitive advantage over China. It’s staggering how rapidly frontier models are growing. As The New York Times reported recently:
"In 2026, GPT 6 Astra took nearly an octillion calculations to train—over 50 million times as many calculations as GPT 1 did. Even calling this figure astronomical understates its size: An octillion is a thousand times the number of stars in the known universe…Researchers predict that by 2030, training the largest A.I. models may use electricity at a rate similar to New York City today."
The bigger models get, the more electricity they demand and therefore, the bigger China’s advantage becomes given their electricity supply capacity and pricing. According to Wall Street Journal reporting from late last year, Chinese data center power rates range between 3 and 5.6¢ per kWh while US data center power rates range between 7 and 9.1¢ per kWh.
Far more to come in the report!
FT link: ft.com/content/e1dd8bff-b06d…
NetAskari retweeted
As #XiJinping again says he wants large numbers of international students in #China, consider this @CDT report: “Twenty Harbin Students Disciplined for Unauthorized VPN Use After Reportedly Being ‘Informed on’ by Classmate” chinadigitaltimes.net/?p=705… via cdt
We are going to have a couple of big stories coming up on China's use of Big Data and AI systems to keep an eye on its own population but also elements abroad.
If you think that the surveillance-state stops at China's borders, you are gravely mistaken.
What once was a practical limitation, only punctuated via dedicated espionage actions in specific cases, has now grown into a full-fledged global drag net. So, stay tuned...
