Please confirm you are human

This browser or connection looks automated. Press and continuously hold the control for 3 seconds to enable Google-hosted web results and, when separately allowed, AI-assisted answers.

A successful check enables 100 search requests. Interactive access does not authorize scraping, systematic collection, or reuse of search output.

Hold with a pointer, or hold Space or Enter.

News

Tech Times
techtimes.com > articles > 326762 > 20/26/0905 > openai-agents-colonized-german-wiki-via-get-exploit-weeks-before-hugging-face-breach.htm

OpenAI Agents Colonized German Wiki Via GET Exploit Weeks Before Hugging Face Breach

1+ week, 19+ hour ago   (298+ words) The agents were running multi-round timed web-lookup tasks — information-retrieval exercises structured as five sequential questions, with a generous first window and extremely short answer deadlines (sometimes 13 seconds) for the rounds that followed. Their environment gave them read access to the…...

Tech Times
techtimes.com > articles > 327148 > 20/26/0910 > magic-matched-deepseek-v4-pro-base-quality-500k-using-50-times-less-compute.htm

Magic Matched DeepSeek V4 Pro Base Quality for $500K, Using 50 Times Less Compute

2+ day, 20+ hour ago   (483+ words) DeepSeek V4 Pro, which launched in April 2026 as an open-weight model with 1.6 trillion total parameters and roughly 49 billion active per inference pass, was itself considered unusually compute-efficient for the quality it delivered. Magic's claimed compute for matching it — approximately 1.58 × 10²³ FLOPs — is…...

Tech Times
techtimes.com > articles > 325581 > 20/26/0826 > single-webpage-can-permanently-poison-local-ai-model-via-nemoclaw-flaw-windows-unpatched.htm

Single Webpage Can Permanently Poison Local AI Model via NemoClaw Flaw, Windows Unpatched

2+ week, 3+ day ago   (515+ words) "The client cannot detect or prevent this — the template is a model-level property invisible to API consumers," Oasis Security stated in its research report. The attack proceeds in three steps that chain together a decades-old browser security technique with a…...

Tech Times
techtimes.com > articles > 326373 > 20/26/0902 > qwen38-max-0902-beats-claude-opus-5-coding-china-law-still-owns-every-call.htm

Qwen3.8-Max-0902 Beats Claude Opus 5 on Coding: China Law Still Owns Every Call

1+ week, 3+ day ago   (382+ words) When Qwen3.8-Max launched on August 3, Arena ranked Qwen3.8-Max fourth on the Code Arena: WebDev leaderboard with 1,668 points — trailing Claude Opus 5 (Max) at 1,705, Kimi K3 (Max) at 1,676, and Claude Opus 5 (High) at 1,669. The strength held across categories. Qwen3.8-Max-0902 ranked first in Data…...

Tech Times
techtimes.com > articles > 326136 > 20/26/0901 > summarizing-website-claude-code-auto-mode-can-compromise-your-machine-no-fix-planned.htm

Summarizing Website in Claude Code Auto Mode Can Compromise Your Machine; No Fix Planned

1+ week, 4+ day ago   (680+ words) Boris Cherny of the Claude Code team said publicly that layered defenses had reduced prompt injection on unseen attacks to approximately zero and that, in practice, Claude Code cannot demonstrate prompt injection anymore. Rehberger's work is now a demonstration. The…...

Tech Times
techtimes.com > articles > 324510 > 20/26/0814 > ai-feature-labels-geometry-not-text-tsinghua-posts-saeverbalizer-preprint.htm

AI Feature Labels From Geometry, Not Text: Tsinghua Posts SAEVerbalizer Preprint

4+ week, 1+ day ago   (344+ words) SAEVerbalizer takes a different approach, working from the inside of the model rather than from observations of its behavior. An SAE produces, for each of its learned features, a decoder direction — a vector in the language model's activation space representing…...

Tech Times
techtimes.com > articles > 325345 > 20/26/0824 > openai-assistants-api-shuts-down-tuesday-no-automated-migration-threads-risk.htm

OpenAI Assistants API Shuts Down Tuesday: No Automated Migration, Threads at Risk

2+ week, 5+ day ago   (368+ words) For any application making calls to /v1/assistants, /v1/threads, or /v1/runs after Tuesday, the response will not be a deprecation warning or a rate limit. It will be a hard error. According to the Assistants API migration guide, OpenAI has provided…...

Tech Times
techtimes.com > articles > 327163 > 20/26/0910 > deepseek-v41-flash-cuts-agent-memory-costs-fourfold-new-architecture.htm

DeepSeek V4.1-Flash Cuts Agent Memory Costs Fourfold With New Architecture

2+ day, 20+ hour ago   (524+ words) The result is a global KV cache of 890 bytes per token — roughly one-quarter of V4-Flash and approximately 1/437th of DeepSeek-V1's per-token global KV cache size from just two years ago. The memory savings come from four distinct architectural changes working…...

Tech Times
techtimes.com > articles > 326941 > 20/26/0908 > liquid-network-exploit-drained-316m-via-software-bug-not-stolen-keys.htm

Liquid Network Exploit Drained $316M Via Software Bug, Not Stolen Keys

4+ day, 17+ hour ago   (619+ words) The incident is the largest publicly disclosed Bitcoin sidechain security breach on record: approximately 3,996 BTC, worth roughly $316 million, drained from the Liquid Federation's reserve wallet in a single transaction, recovered in part within about 26 hours through an entirely on-chain negotiation....

Tech Times
techtimes.com > articles > 326935 > 20/26/0908 > adobe-patches-magento-zero-day-that-hacked-stores-all-patches-applied.htm

Adobe Patches Magento Zero-Day That Hacked Stores With All Patches Applied

4+ day, 18+ hour ago   (572+ words) StyleSmuggler is a two-stage server-side template injection attack (CWE-1336) that exploits Magento's template rendering engine through a specific input property called styles. Server-side template injection is a class of vulnerability in which user-supplied input triggers remote code execution — a distinction…...

Web

External web results are waiting for the human check. Complete the press-and-hold control above. Google advertising and AI choices remain separate after verification.