https://themodelwire.com/article/apple-sues-former-employee-over-alleged-openai-data-theft-and-evidence-destructi-01M1D5BEQD6PQ2QXRK2HNHC274 Modelwire en 2026-09-01T00:13:00.000Z Apple sues former employee over alleged OpenAI data theft and evidence destruction https://themodelwire.com/article/pentagon-consolidates-chatgpt-grok-and-gemini-on-unified-ai-portal-01M1CQM02VB6QEAXC7AY1NKGB9 Modelwire en 2026-08-31T20:13:45.000Z Pentagon consolidates ChatGPT, Grok, and Gemini on unified AI portal https://themodelwire.com/article/instagram-restricts-reach-of-undisclosed-ai-profiles-01M1CM64CZRSS1NK0WHNM37W27 Modelwire en 2026-08-31T19:16:59.000Z Instagram restricts reach of undisclosed AI profiles https://themodelwire.com/article/blue-voice-raises-6m-to-deploy-llm-guidance-for-police-officers-01M1CJF7JMM7X4775RVDHWS383 Modelwire en 2026-08-31T18:35:11.000Z Blue Voice raises $6M to deploy LLM guidance for police officers https://themodelwire.com/article/bank-of-england-flags-ai-valuation-bubble-as-systemic-financial-risk-01M1CFWW6J7VZTPYCKPWCPMV4H Modelwire en 2026-08-31T18:01:21.000Z Bank of England flags AI valuation bubble as systemic financial risk https://themodelwire.com/article/openai-agents-breach-hugging-face-in-sandbox-escape-incident-01M1CGR8RTGD003D2D8JQR36D9 Modelwire en 2026-08-31T18:00:00.000Z OpenAI agents breach Hugging Face in sandbox escape incident https://themodelwire.com/article/decentralized-learning-algorithm-eliminates-horizon-dependent-regret-in-multi-ag-01M1DK2ZJGT178WR380W80DQB4 Modelwire en 2026-08-31T17:59:16.000Z Decentralized learning algorithm eliminates horizon-dependent regret in multi-agent games https://themodelwire.com/article/hybrid-verification-layer-grounds-llm-clinical-screening-in-auditable-guidelines-01M1DGGJN6Z3HEEGMKQ99RB5R0 Modelwire en 2026-08-31T17:40:43.000Z Hybrid verification layer grounds LLM clinical screening in auditable guidelines https://themodelwire.com/article/papergym-enables-rl-training-for-ai-research-planning-without-ground-truth-01M1DGGJP9GP8HF1HA7FM17ZKP Modelwire en 2026-08-31T17:31:18.000Z PaperGym enables RL training for AI research planning without ground truth https://themodelwire.com/article/meta-throttles-unlabeled-ai-profiles-after-admitting-user-confusion-01M1CE5ZMRWA5RPDRSA85Z4HYP Modelwire en 2026-08-31T17:27:43.000Z Meta throttles unlabeled AI profiles after admitting user confusion https://themodelwire.com/article/aspire-benchmark-tests-models-ability-to-self-improve-from-vague-goals-01M1DGGJPXNEATC8G3710QV4A8 Modelwire en 2026-08-31T17:14:59.000Z ASPIRE benchmark tests models' ability to self-improve from vague goals https://themodelwire.com/article/efficient-ai-evaluation-masks-bias-shifts-study-finds-01M1DK2ZPRDJ46GPSJ2M68SW2P Modelwire en 2026-08-31T17:13:15.000Z Efficient AI evaluation masks bias shifts, study finds https://themodelwire.com/article/automated-auditing-framework-surfaces-rare-llm-failures-at-deployment-scale-01M1DGGJQNW38QVZDSJZYTS16M Modelwire en 2026-08-31T17:10:59.000Z Automated auditing framework surfaces rare LLM failures at deployment scale https://themodelwire.com/article/post-training-llms-as-maintained-systems-not-research-artifacts-01M1DK2ZQX162GG5TXBDCV3QW8 Modelwire en 2026-08-31T17:08:41.000Z Post-training LLMs as maintained systems, not research artifacts https://themodelwire.com/article/benchmark-tests-whether-llms-can-self-improve-through-autonomous-testing-and-eva-01M1DGGJR9ACGV6YKHZS0KNTJ9 Modelwire en 2026-08-31T17:05:41.000Z Benchmark tests whether LLMs can self-improve through autonomous testing and evaluation https://themodelwire.com/article/adaptive-data-structuring-cuts-llm-agent-costs-by-28x-on-reasoning-tasks-01M1DGGJSHE1RFZA7AGGM1G77G Modelwire en 2026-08-31T16:53:45.000Z Adaptive data structuring cuts LLM agent costs by 28X on reasoning tasks https://themodelwire.com/article/dpo-training-transfers-sycophancy-from-teacher-to-student-models-01M1DK2ZT02991JA3R9AGFS1SD Modelwire en 2026-08-31T16:52:57.000Z DPO training transfers sycophancy from teacher to student models https://themodelwire.com/article/autoscirub-builds-evaluation-rubrics-before-research-agents-execute-01M1DGGJT6TRMEG7HVGPMD4XMY Modelwire en 2026-08-31T16:48:51.000Z AutoSciRub builds evaluation rubrics before research agents execute https://themodelwire.com/article/reasoning-in-embedding-space-cuts-llm-inference-cost-01M1DGGJTTR1K7KKRN56A93D58 Modelwire en 2026-08-31T16:45:41.000Z Reasoning in embedding space cuts LLM inference cost https://themodelwire.com/article/tiny-transformer-learns-perfect-length-generalization-on-circuit-tasks-01M1DK2ZVNXGHXBVHS7EJ83ZED Modelwire en 2026-08-31T16:42:11.000Z Tiny transformer learns perfect length generalization on circuit tasks https://themodelwire.com/article/prompt-optimization-rivals-model-scaling-for-information-extraction-01M1DGGJW1ZH6BV6JT2M0927SJ Modelwire en 2026-08-31T16:35:31.000Z Prompt optimization rivals model scaling for information extraction https://themodelwire.com/article/on-policy-distillation-gains-dont-come-from-distillation-study-finds-01M1DGGJWCQHX47RBDY1NGYAS3 Modelwire en 2026-08-31T16:22:10.000Z On-policy distillation gains don't come from distillation, study finds https://themodelwire.com/article/normalized-lora-improves-convergence-across-pretraining-and-finetuning-01M1DK2ZX2AD9T32W08DBRB0M3 Modelwire en 2026-08-31T16:15:36.000Z Normalized LoRA improves convergence across pretraining and finetuning https://themodelwire.com/article/autonomous-driving-models-rely-on-memory-more-than-perception-research-finds-01M1DK2ZXRH17FFPTHYZTPQERB Modelwire en 2026-08-31T16:10:04.000Z Autonomous driving models rely on memory more than perception, research finds https://themodelwire.com/article/one-in-three-clinical-ai-notes-contain-verified-errors-audit-finds-01M1DGGJY21VBTKP39SP08SPZX Modelwire en 2026-08-31T16:00:27.000Z One in three clinical AI notes contain verified errors, audit finds https://themodelwire.com/article/clipto-reaches-profitability-and-250m-valuation-on-video-search-ai-01M1C9WJPGGP4F22EEWYCQ7FA4 Modelwire en 2026-08-31T16:00:00.000Z Clipto reaches profitability and $250M valuation on video search AI https://themodelwire.com/article/llm-judges-fail-to-catch-omissions-in-ai-drafted-clinical-notes-01M1DGGJYJ36KMKFK89WCNTNW5 Modelwire en 2026-08-31T15:59:54.000Z LLM judges fail to catch omissions in AI-drafted clinical notes https://themodelwire.com/article/language-guided-flow-matching-accelerates-structure-based-molecular-generation-01M1DK300TMYPZ8JY0FDT32JGA Modelwire en 2026-08-31T15:57:09.000Z Language-guided flow matching accelerates structure-based molecular generation https://themodelwire.com/article/quantized-caches-break-rag-faithfulness-despite-maintaining-accuracy-01M1DGGK0FDERH0276T39H0FTY Modelwire en 2026-08-31T15:47:23.000Z Quantized caches break RAG faithfulness despite maintaining accuracy https://themodelwire.com/article/llms-fail-basic-self-awareness-tests-synthetic-training-shows-limited-gains-01M1DGGK28KNNSKFGTJXTPRC24 Modelwire en 2026-08-31T15:37:51.000Z LLMs fail basic self-awareness tests, synthetic training shows limited gains https://themodelwire.com/article/cogevol-replaces-multi-turn-agents-with-single-pass-learning-content-generation-01M1DGGK2WQZ95NYH1VD2HS5Y2 Modelwire en 2026-08-31T15:33:38.000Z CogEvol replaces multi-turn agents with single-pass learning content generation https://themodelwire.com/article/cross-model-kv-cache-sharing-cuts-redundant-prefill-computation-01M1DK3035HCXXWMR25R3JABFK Modelwire en 2026-08-31T15:28:17.000Z Cross-model KV cache sharing cuts redundant prefill computation https://themodelwire.com/article/openais-ad-business-reaches-1-billion-annual-run-rate-01M1C7A7X6WRRJTK5D7V7FW4C8 Modelwire en 2026-08-31T15:24:59.000Z OpenAI's ad business reaches $1 billion annual run rate https://themodelwire.com/article/cxmt-begins-hbm3e-production-reducing-chinas-ai-memory-dependency-01M1C6ES36JTB761TD89944ZNG Modelwire en 2026-08-31T15:17:31.000Z CXMT begins HBM3E production, reducing China's AI memory dependency https://themodelwire.com/article/nvidia-bets-3-5b-on-mediatek-as-hyperscalers-build-rival-ai-chips-01M1C6EPBMT9CFTC6BX9GY83BC Modelwire en 2026-08-31T15:15:25.000Z Nvidia bets $3.5B on MediaTek as hyperscalers build rival AI chips https://themodelwire.com/article/eu-classifies-chatgpt-as-very-large-search-engine-under-digital-services-act-01M1C3WBYW71M9SMGM2TPC6BRP Modelwire en 2026-08-31T14:31:37.000Z EU classifies ChatGPT as very large search engine under Digital Services Act https://themodelwire.com/article/openai-moves-to-outcome-based-pricing-with-enterprise-customers-01M1C3WBZCYNHXNF9SYG0C2ZQV Modelwire en 2026-08-31T14:28:14.000Z OpenAI moves to outcome-based pricing with enterprise customers https://themodelwire.com/article/hochul-stakes-out-aggressive-ai-regulation-stance-in-election-year-01M1C25C5GCNXKHPTW3AGT4MGJ Modelwire en 2026-08-31T14:00:00.000Z Hochul stakes out aggressive AI regulation stance in election year https://themodelwire.com/article/anthropic-launches-hardware-standard-for-ai-agent-machine-control-01M1C25E9CKYNBMM2VV6MA2CGZ Modelwire en 2026-08-31T13:50:28.000Z Anthropic launches hardware standard for AI agent machine control https://themodelwire.com/article/eu-classifies-chatgpt-as-major-online-platform-triggering-stricter-oversight-01M1C0EE2CFTJNYB8S079D323R Modelwire en 2026-08-31T13:27:36.000Z EU classifies ChatGPT as major online platform, triggering stricter oversight https://themodelwire.com/article/instagram-tightens-detection-of-fake-ai-influencers-relabels-disclosure-01M1BYQHADZB4SH3JBSXB1CPCT Modelwire en 2026-08-31T13:00:00.000Z Instagram tightens detection of fake AI influencers, relabels disclosure https://themodelwire.com/article/swarmbench-measures-llm-orchestration-gaps-in-multi-agent-systems-01M1DC792BZ57R9CG3XQH85AQF Modelwire en 2026-08-31T12:02:04.000Z SwarmBench measures LLM orchestration gaps in multi-agent systems https://themodelwire.com/article/new-benchmark-exposes-multimodal-model-hallucinations-across-complex-image-reaso-01M1DESMSVATXKBCE170WEN7PZ Modelwire en 2026-08-31T11:53:03.000Z New benchmark exposes multimodal model hallucinations across complex image reasoning https://themodelwire.com/article/geometric-signals-reveal-when-llm-reasoning-destabilizes-across-long-conversatio-01M1DC7931D9J60TXA0DXJX2WQ Modelwire en 2026-08-31T11:51:54.000Z Geometric signals reveal when LLM reasoning destabilizes across long conversations https://themodelwire.com/article/researchers-pinpoint-identity-storage-in-multimodal-models-enable-targeted-remov-01M1DC793NFR8DJPF9ARMDW97T Modelwire en 2026-08-31T11:51:05.000Z Researchers pinpoint identity storage in multimodal models, enable targeted removal https://themodelwire.com/article/perplexity-guided-reasoning-annotations-scale-pretraining-data-synthesis-01M1DC7983F2V132G4GYT9EHPC Modelwire en 2026-08-31T11:34:19.000Z Perplexity-guided reasoning annotations scale pretraining data synthesis https://themodelwire.com/article/aligned-models-vulnerable-to-bias-injection-via-benign-synthetic-data-01M1DC799FESYXH06MMK7JXX8S Modelwire en 2026-08-31T11:28:36.000Z Aligned models vulnerable to bias injection via benign synthetic data https://themodelwire.com/article/mechanistic-analysis-reveals-how-roleplay-wrappers-disable-llm-safety-refusals-01M1DESN0PWPHARGDN89E6D4HG Modelwire en 2026-08-31T11:03:49.000Z Mechanistic analysis reveals how roleplay wrappers disable LLM safety refusals