https://www.seventnews.com/en/articles/claude-published-malware-to-pypi-because-it-thought-the-internet-was-fake SevenTnewS en 2026-08-02T19:26:57.753988+02:00 Claude published malware to PyPI because it thought the internet was fake https://www.seventnews.com/en/articles/cisa-adds-two-actively-exploited-flaws-to-its-must-patch-catalog SevenTnewS en 2026-08-02T18:30:00+02:00 CISA adds two actively exploited flaws to its must-patch catalog https://www.seventnews.com/en/articles/why-your-ai-tutor-cant-see-how-math-builds-on-itself SevenTnewS en 2026-08-02T17:00:00+02:00 Why your AI tutor can't see how math builds on itself https://www.seventnews.com/en/articles/957253-records-cant-hide-the-gap-ai-agents-surge-in-coding-but-stall-where-enterprises-need-them SevenTnewS en 2026-08-02T16:00:00+02:00 957,253 records can't hide the gap: AI agents surge in coding but stall where enterprises need them https://www.seventnews.com/en/articles/5-million-votes-25-elo-points-inside-the-chatbot-arena-leaderboard-nobody-can-shake SevenTnewS en 2026-08-02T15:00:00+02:00 5 Million Votes, 25 Elo Points: Inside the Chatbot Arena Leaderboard Nobody Can Shake https://www.seventnews.com/en/articles/the-tax-loophole-coinbase-is-using-to-sell-you-leveraged-crypto-futures SevenTnewS en 2026-08-02T13:30:00+02:00 The tax loophole Coinbase is using to sell you leveraged crypto futures https://www.seventnews.com/en/articles/neuro-symbolic-reasoning-outperforms-raw-neural-models-on-temporal-tasks-benchmark-finds SevenTnewS en 2026-08-02T13:00:00+02:00 Neuro-symbolic reasoning outperforms raw neural models on temporal tasks, benchmark finds https://www.seventnews.com/en/articles/a-plastic-e-reader-case-just-broke-the-logic-of-gun-blocking-software SevenTnewS en 2026-08-02T12:00:00+02:00 A plastic e-reader case just broke the logic of gun-blocking software https://www.seventnews.com/en/articles/sparking-zeros-rpg-progression-mode-redefines-what-fighting-game-dlc-can-be SevenTnewS en 2026-08-02T11:00:00+02:00 Sparking Zero's RPG progression mode redefines what fighting game DLC can be https://www.seventnews.com/en/articles/on-swe-bench-verified-top-models-hit-96-on-private-enterprise-code-they-barely-clear-23 SevenTnewS en 2026-08-02T11:00:00+02:00 On SWE-bench Verified, Top Models Hit 96%. On Private Enterprise Code, They Barely Clear 23% https://www.seventnews.com/en/articles/bending-spoons-playbook-buy-old-tech-make-it-new-and-bet-on-refreshing SevenTnewS en 2026-08-02T09:00:00+02:00 Bending Spoons' playbook: Buy old tech, make it new, and bet on 'refreshing' https://www.seventnews.com/en/articles/research-idea-generation-gets-389x-better-with-ideagent SevenTnewS en 2026-08-02T09:00:00+02:00 Research idea generation gets 3.89x better with IDEAgent https://www.seventnews.com/en/articles/ai2s-olmoearth-platform-turns-satellite-inference-into-a-distributed-systems-problem SevenTnewS en 2026-08-02T08:00:00+02:00 Ai2's OlmoEarth platform turns satellite inference into a distributed systems problem https://www.seventnews.com/en/articles/cisco-flaw-on-cisas-kev-list-patching-is-not-enough-anymore SevenTnewS en 2026-08-02T08:00:00+02:00 Cisco flaw on CISA's KEV list: patching is not enough anymore https://www.seventnews.com/en/articles/the-weekly-screenshot-contest-that-keeps-black-flag-resynced-players-coming-back SevenTnewS en 2026-08-01T18:30:00+02:00 The weekly screenshot contest that keeps Black Flag Resynced players coming back https://www.seventnews.com/en/articles/for-agentic-ai-per-call-routing-is-the-blind-spot-trace-router-has-a-fix SevenTnewS en 2026-08-01T17:00:00+02:00 For agentic AI, per-call routing is the blind spot. TRACE-Router has a fix. https://www.seventnews.com/en/articles/ball-x-pits-final-free-update-adds-2-new-characters-11-balls-and-a-decision-between-firepower-and-backpack-space SevenTnewS en 2026-08-01T16:30:00+02:00 Ball x Pit's Naturalist Update: 11 balls, 2 characters, one impossible choice https://www.seventnews.com/en/articles/the-ai-that-promises-to-clear-80-of-your-security-backlog SevenTnewS en 2026-08-01T16:00:00+02:00 Devin-powered service promises to clear 80% of security backlogs https://www.seventnews.com/en/articles/grok-45s-biggest-news-might-be-how-it-was-built SevenTnewS en 2026-08-01T14:30:00+02:00 Grok 4.5's biggest news might be how it was built https://www.seventnews.com/en/articles/agentrca-zero-shot-root-cause-analysis-that-explains-its-own-reasoning SevenTnewS en 2026-08-01T13:00:00+02:00 AgentRCA: zero-shot root cause analysis that explains its own reasoning https://www.seventnews.com/en/articles/cursors-india-plan-at-649-inr-per-month-includes-a-model-router-that-picks-the-cheapest-model-for-each-task SevenTnewS en 2026-08-01T12:30:00+02:00 Cursor's India plan: model router picks cheapest AI, 649 INR/month https://www.seventnews.com/en/articles/training-surgical-robots-just-got-faster-nvidias-cosmos-h-dreams-runs-at-160-fps SevenTnewS en 2026-08-01T12:00:00+02:00 Training surgical robots just got faster: Nvidia's Cosmos-H-Dreams runs at 160 fps https://www.seventnews.com/en/articles/gpqa-diamond-was-built-to-be-google-proof-frontier-models-are-now-clearing-95-anyway SevenTnewS en 2026-08-01T11:00:00+02:00 GPQA Diamond Was Built to Be Google-Proof. Frontier Models Are Now Clearing 95% Anyway https://www.seventnews.com/en/articles/zero-shot-transfer-on-graphs-has-a-multimodal-blind-spot-charm-offers-a-fix SevenTnewS en 2026-08-01T10:30:00+02:00 Zero-shot transfer on graphs has a multimodal blind spot. CHARM offers a fix https://www.seventnews.com/en/articles/openforgerl-trains-ai-agents-without-touching-their-inference-harnesses SevenTnewS en 2026-08-01T09:00:00+02:00 OpenForgeRL trains AI agents without touching their inference harnesses https://www.seventnews.com/en/articles/anthropics-claude-finds-its-enterprise-bridge-30000-cognizant-engineers SevenTnewS en 2026-08-01T08:00:00+02:00 Anthropic's Claude finds its enterprise bridge: 30,000 Cognizant engineers https://www.seventnews.com/en/articles/ai-desktop-agents-fail-before-after-test-35-of-the-time SevenTnewS en 2026-08-01T08:00:00+02:00 AI desktop agents fail before-after test 35% of the time https://www.seventnews.com/en/articles/the-ai-pipeline-has-three-silent-leaks-and-every-fix-makes-one-of-them-worse SevenTnewS en 2026-07-31T21:00:00+02:00 The AI Pipeline Has Three Silent Leaks, and Every Fix Makes One of Them Worse