Chinese labs reportedly distill OpenAI GPT-6 Astra outputs
The Information says DeepSeek and Moonshot used Astra, including hidden reasoning, to train Kimi K3.1 Pro and V4.1-Flash 1017.
62
Organization
DeepSeek
- Mentions · 7d2
- Mentions · 30d22
- All mentions56
- First seenAug 31, 2026
- Last mentioned1d ago
Timeline
56 articles · newest firstOpenAI patches compaction smuggling used to steal model reasoning
Xint received a $6,000 bounty on October 23 after OpenAI confirmed and fixed the bypass of its August mitigations.
64
DeepSeek reportedly raises about $15 billion after US curbs
Bloomberg says the round is about 100 billion yuan, with CATL and Tencent among potential investors.
68
DeepSeek open-sources DeepInfer, its production inference engine
DeepSeek claims 155 trillion tokens a day and 96% cache hits; KV cache and load balancing are not fully released.
62
web-edit 0.1a2
Simon Willison shipped web-edit 0.1a2 with clipboard copy while testing DeepSeek 4.1 Flash pelican prompts.
16
Deepseek V4.1 Flash 1017 beats larger rivals in agentic tasks while once again massively cutting per-task costs
DeepSeek shipped open-weight V4.1-Flash-1017, claiming better agentic coding at under $0.20 per task.
73
DeepSeek Just Released V4.1-Flash-1017, Scores 77.0 on DeepSWE, Opens Search API
DeepSeek released V4.1-Flash-1017, scoring 77.0 on DeepSWE and opening a Search API.
64
DeepSeek: DeepSeek v4.1 Flash 1017
DeepSeek's October 17 V4.1 Flash snapshot is a 552B MoE with native vision and 1M context.
74
DeepSeek-v4.1-Flash-1017
DeepSeek posted DeepSeek-V4.1-Flash-1017 on Hugging Face, drawing early Hacker News discussion.
46
Conditional Program stores LLM n-gram lookup tables at two bits
For DeepSeek, the method cuts table storage from 203GB to 51GB versus keeping full embeddings.
44
GPT-6-Astra solves 49 of 50 physics research problems at $2.40 apiece
OpenAI's GPT-6-Astra solved 49 of 50 routine theoretical-physics problems at about $2.40 each.
58
Anthropic Warns Small AI Models Could Soon Form Self-Propagating IoT Botnets
Anthropic warns small open-weight models may soon self-replicate across insecure IoT devices as AI botnets.
68
Solo Hacker Used AI Tools to Breach South Korean Financial Organizations and Steal Data
Suspected lone Chinese-speaking hacker used ARTEX AI pentest tool to breach Shinhan, Kookmin, Hana and other South Korean financial firms, stealing customer data.
74
A Closer Look at Agentic BBO: Benchmarking LLM Agents for Black-Box Optimization
AgenticBBO-Bench compares LLM agents on black-box optimization across five domains, with GPT-6 Astra on the cost-performance frontier.
50
EngramEdit: Decoupled Knowledge Updates in LLMs through Conditional Memory
EngramEdit updates LLM facts in conditional memory while keeping the Transformer backbone fixed.
52
Introducing Mistral Large 4: Le chonk
Mistral previewed Mistral Large 4, a 1-trillion-parameter model with 49 billion active parameters.
76
EngramEdit: Decoupled Knowledge Updates in LLMs through Conditional Memory
EngramEdit updates facts in DeepSeek Engram conditional memory while keeping the Transformer backbone fixed.
54
CATL and Tencent back Deepseek's ballooning funding round as the AI startup eyes a 2027 IPO
DeepSeek is nearing a $12–15 billion round backed by CATL and Tencent, ahead of a planned early 2027 IPO.
76
DeepSeek Harness v0.2 Brings Official Desktop Apps to Its Open-Source Agent Harness
DeepSeek released official macOS and Windows desktop apps for MIT-licensed DeepSeek Harness v0.2.
46
[AINews] not much happened today
OpenAI’s GPT-6.1 Sol undercuts Astra on price while Sonnet 5.5 ranks high on Agent Arena.
80
US arrests tech CEO accused of smuggling $300M in Nvidia chips into China
The US arrested Earthmade CEO Greg Lui for allegedly smuggling over $300 million in Nvidia AI chips to China.
78
DeepSeek Harness
DeepSeek released DeepSeek Harness, an open-source agent framework now in worldwide public preview.
58
The AI Race Just Got Awkward
A commentary says Western labs adopted DeepSeek KV-cache optimizations and cut cache-read prices.
34
ID Balancing: Stable Training of Extremely Sparse MoE via PID-Based Load Control
ID Balancing uses an integral-derivative controller to stabilize extremely sparse MoE expert load during training.
50
Will Chinese AI companies slow down? A top House Democrat wants answers
Rep. Ro Khanna urges a US-China treaty banning recursive self-improvement, writing to ODNI and Chinese AI labs DeepSeek, Alibaba, and Moonshot AI.
52
Social Engineering in the Age of Synthetic Media
Recorded Future details how threat actors abuse jailbroken and legitimate AI tools to scale phishing, deepfake impersonation, and phishing-as-a-service operations.
55
Predictive Credit: Measuring What Scientific Explanations Add to Experimental Forecasts
A protocol finds natural research-agent explanations do not yet show confirmed predictive credit on Tox21 and OpenML.
34
DeepSeek Elastic Compute (DSec)
DeepSeek’s arXiv preprint on Elastic Compute (DSec) drew a short Hacker News discussion.
36
AgentWorld: Benchmarking Long-Horizon Collaboration of Multi-agent LLMs
AgentWorld finds even the best tested LLM succeeds on only 52 percent of long-horizon multi-agent tasks.
58
Prefilling the Reasoning Channel: Output-Prefix Attacks on Reasoning LLMs
Output-prefix attacks plus injected reasoning jailbreak Gemini, DeepSeek, and Claude models at up to 99% success.
62
This Malware Doesn’t Wait for Hackers—It Asks AI Models What to Do Next
Cisco Talos documents CLOSEDQUORUM, a Windows implant that votes across DeepSeek, Qwen, Mistral, and Gemini APIs to choose attack actions autonomously.
58
AI malware just removed the human from the attack loop
Cisco Talos uncovers CLOSEDQUORUM, first 'LLM-as-C2' malware using multi-model judge panel to autonomously steal credentials and crypto wallets.
62
Windows CLOSEDQUORUM malware uses AI models to autonomously select post-compromise actions
Cisco Talos documents CLOSEDQUORUM, the first known Windows implant using a four-LLM quorum to autonomously choose credential theft, wallet theft, injection, and persistence actions.
60
Autonomous AI Agents Hack Retailers for $25 and Steal 600,000 Credit Cards
Gambit Security details an unsupervised AI-agent campaign using Strix, Cairn and Hermes that breached 27+ retailers and stole 600,000+ credit cards for about $25 per target.
78
The current balance of power in open models
A report finds Chinese open-weight AI models now outperform American counterparts in downloads and benchmarks.
45
DeepSeek-v4.1 Flash: Pushing the Limits of KV Cache Compression
DeepSeek-V4.1 Flash is a 552B-parameter multimodal MoE model with 1M-token context achieving 4x KV cache compression for long-horizon agent workloads.
74
DeepSeek-V4.1-Flash: Pushing the Limits of KV Cache Compression
DeepSeek releases V4.1-Flash, a 552B-parameter multimodal MoE with 1M-token context and 890-byte-per-token KV cache for cheaper agentic workloads; weights on Hugging Face.
84
Show HN: How Stale Is Your AI? Release age and training cutoff for 20 models
A new site tracks release dates and training cutoffs for 20 AI models across 8 labs, exposing months-long staleness gaps.
12
[AINews] DeepSeek v4.1-Flash: 763B-P8B-D16B novel causal Encoder–Decoder architecture with vision marks the Return of the Whale
DeepSeek released V4.1-Flash, an open-weight 763B-parameter model with a novel causal encoder-decoder architecture, 1M context, vision input, and MIT license.
80
Anthropic Says Seven China-Based AI Labs Ran Industrial-Scale Claude Distillation Attacks
Anthropic disrupted industrial-scale unauthorized Claude distillation by seven China-based AI labs, including Alibaba, DeepSeek, Moonshot, and Z.ai.
76
Related CVEs
by mentions- CVE-2026-8207861%9.4Unsafe Reflection RCE in PaperCut NG/MF, Chained with Auth Bypass in Attacks
- CVE-2026-8157884%8.8Missing Authentication for Critical Function in PaperCut NG/MF Web Interface
- CVE-2021-4228777%7.5Privilege Escalation in Microsoft Active Directory Domain Services
- CVE-2021-4227873%7.5Privilege Escalation via sAMAccountName Spoofing in Microsoft Active Directory
Appears with
- DeepSeekvendor · 21
- Moonshot AIorganization · 15
- Anthropicorganization · 15
- Alibabaorganization · 13
- Anthropicvendor · 13
- OpenAIorganization · 13
- OpenAIvendor · 12
- Claudeai model · 11
- Claudeproduct · 9
- MiniMaxorganization · 8
- Qwenai model · 8
- Googleorganization · 8
- Z.aiorganization · 8
- FBIorganization · 7
- Alibabavendor · 7
- Geminiai model · 7
Entities are extracted by the model from each article. First seen , last mentioned . Watching keeps it in this browser only (no account); the watchlist page and dashboard alerts use it.