https://hashsparks.org/stories/ai-assisted-gpu-porting-cress-weather-model/HashSparksen2026-08-16The Most Important Part of This AI-Assisted GPU Port Was the Test Harnesshttps://hashsparks.org/stories/anthropic-multiagent-swarms-collusion-sabotage/HashSparksen2026-08-16Anthropic's Agent Swarms Reported More Findings—and New Ways to Fail Togetherhttps://hashsparks.org/stories/big-pickle-swe-atlas-50-8-self-reported/HashSparksen2026-08-16Big Pickle's 50.8% SWE Atlas Run Is a Self-Reported Result, Not a Rankhttps://hashsparks.org/stories/jitpass-macos-secrets-touch-id-limits/HashSparksen2026-08-16Jit Puts Mac Developer Secrets Behind Its Touch ID Workflow. The Keychain Boundary Is More Complicatedhttps://hashsparks.org/stories/littlelearner-controlled-pretraining/HashSparksen2026-08-16LittleLearner puts a curriculum boundary around pretraining—imperfectly, but usefullyhttps://hashsparks.org/stories/proofrun-local-verification-receipts/HashSparksen2026-08-16ProofRun Records Whether Tests Ran on This Code. That Is Not Proof the Code Is Correcthttps://hashsparks.org/stories/ai-account-security-2026/HashSparksen2026-08-15How to Check for Signs Someone Else Is Using Your ChatGPT, Claude or Gemini Accounthttps://hashsparks.org/stories/ai-designed-bacteriophages-not-human-virus/HashSparksen2026-08-15AI-Generated Genomes Yielded 16 Working Bacteriophages. No Human Virus Was Testedhttps://hashsparks.org/stories/anthropic-august-2026-risk-report-safety-failures/HashSparksen2026-08-15Anthropic revises AI risk assessments as it details safeguard failureshttps://hashsparks.org/stories/brain-building-algorithm-thought-experiment/HashSparksen2026-08-15The Brain-Building Algorithm Is a Thought Experiment, Not a Blueprinthttps://hashsparks.org/stories/debian-llm-vote-2026/HashSparksen2026-08-15Debian Is Voting on Eight Different Futures for AI-Written Contributionshttps://hashsparks.org/stories/disney-line-art-vectorization-2026/HashSparksen2026-08-15Disney and ETH's Line-Art Vectorizer Produces Editable Bézier Strokes—Within Limitshttps://hashsparks.org/stories/gen-z-cinema-comeback-evidence/HashSparksen2026-08-15Young U.S. adults lead in cinema reach. A Gen Z comeback is a bigger claimhttps://hashsparks.org/stories/google-heir-homomorphic-encryption-private-ai/HashSparksen2026-08-15Google’s HEIR Compiles AI for Encrypted Data. “Practical” Still Has an Asteriskhttps://hashsparks.org/stories/gpu-kernel-verifier-llm-correctness-audit/HashSparksen2026-08-15AI-Written GPU Kernels Can Pass the Test and Still Be Wrong, a New Audit Argueshttps://hashsparks.org/stories/lymealert-2026-home-tick-test/HashSparksen2026-08-15LymeAlert’s Home Test Is Designed to Check the Tick—Not the Person Bittenhttps://hashsparks.org/stories/openai-europe-privacy-preview-ads-launch/HashSparksen2026-08-15OpenAI's August privacy preview is not an EEA ad-launch datehttps://hashsparks.org/stories/us-water-utility-cyberattacks-iran-attribution/HashSparksen2026-08-15What the U.S. Has — and Has Not — Said About the Water-Utility Cyberattacks