https://anguishedturtle.com/feed.xml Ranked AI and technology signals Thu, 20 Aug 2026 18:11:19 +0000 https://cisa.gov/news-events/alerts/2026/08/19/cisa-adds-one-known-exploited-vulnerability-catalog sw-8a10b190281ff8214b07 Wed, 19 Aug 2026 12:00:00 +0000 CISA has added one new vulnerability to its Known Exploited Vulnerabilities (KEV) Catalog, based on evidence of active exploitation. CVE-2026-64849 MLflow Server-Side Request Forgery Vulnerability This type of vulnerability is a frequent attack vector for malicious cyber actors and poses significant risks to the federa security https://platform.claude.com/docs/en/release-notes/overview sw-4182359e79e632d3734e Wed, 19 Aug 2026 00:00:00 +0000 The Files API is now generally available on the Claude API. Requests to the /v1/files endpoints, and Messages API requests that reference an uploaded file, no longer require the files-api-2025-04-14 beta header. Requests sent without the header use the GA response format: file expiration (set expires_in_seconds when yo products https://cisa.gov/news-events/alerts/2026/08/18/cisa-adds-four-known-exploited-vulnerabilities-catalog sw-4d476a0816ad5637f4cb Tue, 18 Aug 2026 12:00:00 +0000 CISA has added four new vulnerabilities to its Known Exploited Vulnerabilities (KEV) Catalog, based on evidence of active exploitation. CVE-2026-33824 Microsoft Internet Key Exchange (IKE) Service Extensions Double Free Vulnerability CVE-2026-55040 Microsoft SharePoint Weak Authentication Vulnerability CVE-2026-59310 B security https://cisa.gov/news-events/alerts/2026/08/20/cisa-adds-two-known-exploited-vulnerabilities-catalog sw-fb28dfd94186606c3786 Thu, 20 Aug 2026 12:00:00 +0000 CISA has added two new vulnerabilities to its Known Exploited Vulnerabilities (KEV) Catalog, based on evidence of active exploitation. CVE-2026-72529 TrueConf Server Missing Authentication for Critical Function Vulnerability CVE-2026-72530 TrueConf Server Code Injection Vulnerability These types of vulnerabilities are security https://arxiv.org/abs/2608.19181v1 sw-64bae30ff0bef819129a Wed, 19 Aug 2026 17:54:58 +0000 On-policy distillation (OPD) trains a student on its own responses using dense token-level guidance from a stronger teacher. In long-context tasks, however, token-level teacher support can favor locally plausible responses that omit evidence distributed across the input or violate global task constraints. Task-specific research https://github.com/anthropics/claude-code/releases/tag/v2.1.237 sw-09d0c8e5c7031ff7de10 Thu, 20 Aug 2026 00:54:41 +0000 What's changed Fixed prompt caching for sessions using an LLM gateway or custom base URL Added a built-in "Concise" output style: Claude leads with results and skips preamble and narration, while doing the work just as thoroughly. Select it under Output style in /config. products https://arxiv.org/abs/2608.19098v1 sw-4cec3ce720b6405610d5 Wed, 19 Aug 2026 16:50:39 +0000 Multi-teacher on-policy distillation (M-OPD) has emerged as a promising paradigm for consolidating domain-specialized reinforcement learning (RL) experts into a single generalist student via dense, token-level reward supervision. Despite its practical success, the optimization dynamics governing multi-teacher capabilit research https://arxiv.org/abs/2608.19072v1 sw-354edfe16fa23bfa768a Wed, 19 Aug 2026 16:17:39 +0000 Large language model (LLM) agents can now post-train an LLM end-to-end. They can write code, launch training, evaluate checkpoints, and improve downstream performance, raising the prospect of AI-for-AI. We argue that this picture conflates two distinct capabilities: execution-level capability, iterating within a select research https://openai.com/index/offering-zero-data-retention-for-frontier-models sw-df8f12302a5b421ac186 Wed, 19 Aug 2026 19:00:00 +0000 OpenAI reaffirms Zero Data Retention for eligible API customers and previews Private Safety Processing for advanced AI safety without compromising data privacy. labs https://openrouter.ai/blog/announcements/openrouter-is-joining-stripe sw-8e399e61dadd7f7793d3 Wed, 19 Aug 2026 17:32:38 +0000 Previously: Stripe will reportedly acquire OpenRouter for $7B+ https://news.ycombinator.com/item?id=49323381 developer https://arxiv.org/abs/2608.16753 sw-3033c57db999fcb8d77f Wed, 19 Aug 2026 15:14:25 +0000 developer https://cerebras.ai/cs4 sw-69778c61809e1384aec3 Wed, 19 Aug 2026 00:28:18 +0000 developer https://github.blog/changelog/2026-08-20-windows-11-arm64-vs2026-image-generally-available sw-3501b3426269f2d41da0 Thu, 20 Aug 2026 17:52:38 +0000 The Windows 11 arm64 image with Visual Studio 2026 is now generally available on standard and larger GitHub-hosted runners. To use it in GitHub Actions, update your workflow file to… The post Windows 11 arm64 VS2026 image generally available appeared first on The GitHub Blog. developer https://huggingface.co/blog/LiquidAI/lfm25-dspark sw-eab5285cfdebeb65a8ef Thu, 20 Aug 2026 16:52:57 +0000 labs https://news.google.com/rss/articles/CBMingFBVV95cUxQRG1ONC11bUpzNWFwMGdRbVVUM0ZHb0l0VHJfU0dsYi1KdmxTZVRnRllWaDE2MkZtSFExX0U5MGtGRFpkOGFYZmY1OHJtMEV4UG9MVUlfc3hUQ2dMTlJOZ2d0R3cwZjB4OFkxbW1rNHhaNEt0alp6RGdGVTRJb201WWloTTR3Y2hKVVdzN0paZXFEUmVmQl9VeHFqZ1dyQQ?oc=5 sw-3b6431d618d073c2eee8 Thu, 20 Aug 2026 17:36:35 +0000 EXCLUSIVE: How a Texas student blew the whistle on a rogue AI hacking attempt Reuters industry https://news.google.com/rss/articles/CBMiqAFBVV95cUxOMldWMkxfMnFZbHg4YnI0RTdoZGtfMXJIT1hTdHg4aGUyMWhvazNtdWVoLW4wMDR0cXkybXhEMHdqanBCZ0Nmb21xVVVua2lYOTQ4elZsUFlMbTBVYkpwRDBaVWs5Y0xRTUc1U3BSVng1WlpteXhqS0NhdWZ4cmthTlpFYTJJX2hrdGc0Y1VZUTJ5aXNzQWlvb19nei05RHJIYlhEZXRSYko?oc=5 sw-9c69e2cdf7108e67f52f Thu, 20 Aug 2026 17:34:02 +0000 Nvidia to ship AI chip for China by year-end, The Information reports Reuters industry https://blog.cloudflare.com/task-based-oauth-consent sw-5ee85db9a39ca3d926e2 Thu, 20 Aug 2026 17:03:03 +0000 Cloudflare OAuth now supports optional scopes, giving users more control over what an app can access and helping developers build secure consent flows around the task at hand. developer https://github.blog/changelog/2026-08-20-code-scanning-adds-a-mitigated-alert-dismissal-reason sw-613ce75ddf903076291f Thu, 20 Aug 2026 15:14:31 +0000 You can now dismiss a code scanning alert with the reason Mitigated when a vulnerability remains in the code but external controls, such as a web application firewall or network… The post Code scanning adds a mitigated alert dismissal reason appeared first on The GitHub Blog. developer https://github.blog/changelog/2026-08-20-separate-github-actions-path-for-github-code-quality sw-a4ff12b9ae120e12841e Thu, 20 Aug 2026 14:29:27 +0000 A dedicated workflow path for code quality CodeQL actions workflows is now generally available. Your workflow run history and your Actions usage reports now tell GitHub Code Quality runs apart… The post Separate GitHub Actions path for GitHub Code Quality appeared first on The GitHub Blog. developer https://news.google.com/rss/articles/CBMiyAFBVV95cUxOeENQNnM1RGoxMUFwdGJ2c19GUDRnTUhiWVdieDZNZG1qX2QzN29SYVBqU0NBcnVaMjRISjB3ZFc5WTlRbXYwQ1FkallpcmdLeEcwWk1wcS1pbG9TbTB6WFQ4VENCa25wMTE1TmJMeGJ1dEFtY0xrX0hOb0ZURkZJWlRmM3lFcHFZaXpqTWFCYVhHdDc0UXJzNEpPTUo5LXdBYzlCcXppZklKbkN5cnc4NGFsU3p5U0xPNnJ1dGUxYWEzQVI0Y3ZqcA?oc=5 sw-5ebe63adb857ea708e7f Thu, 20 Aug 2026 16:00:47 +0000 The new kingmakers: Crypto, AI and betting firms fuel record spending on the 2026 midterms Reuters industry https://arxiv.org/abs/2608.18940v1 sw-6cf2ee0e9e4321841eea Wed, 19 Aug 2026 14:08:05 +0000 Single-step retrosynthesis is a central component of computer-aided synthesis planning, yet its intrinsically one-to-many nature is poorly captured by single-answer evaluation and benchmarking protocols. To address this, we introduce Top-K prompting as a robust training and inference paradigm to better capture diverse, research https://arxiv.org/abs/2608.18827v1 sw-f0e1f69eb0834c3c3a5d Wed, 19 Aug 2026 11:36:55 +0000 Reward function design remains a bottleneck in reinforcement learning. While large language models (LLMs) have enabled automated reward generation, existing methods generate and revise reward functions as monolithic programs, making it difficult to reliably preserve and reuse effective components discovered in earlier research https://arxiv.org/abs/2608.18825v1 sw-39c87801de19e9052ef3 Wed, 19 Aug 2026 11:29:30 +0000 Medical automatic speech recognition (MedASR) requires adaptation to specialised terminology, limited annotated clinical data, and multilingual use cases. Although large-scale pretrained ASR models such as Whisper achieve strong generalisation, their behaviour after medical and multilingual adaptation remains insuffici research https://cisa.gov/news-events/cybersecurity-advisories/aa26-231a sw-2c28fe09bfb404db3c29 Wed, 19 Aug 2026 12:00:00 +0000 Executive summary Note: This advisory relates to an active threat to Siemens S7 Series programmable logic controllers (PLCs). However, ongoing PLC targeting activity is broader than Siemens PLCs. All PLC owners and operators should apply relevant mitigations to reduce the risk to their devices and systems. The Siemens- security https://linear.app/data sw-75bb8d6c5587bf665d9f Tue, 18 Aug 2026 22:08:07 +0000 developer https://cisa.gov/news-events/ics-advisories/icsa-26-232-01 sw-83373ca359617e005072 Thu, 20 Aug 2026 12:00:00 +0000 View CSAF Summary Successful exploitation of this vulnerability could allow a local attacker with low privileges to extract user credentials (passwords and authentication tokens) from system memory, potentially leading to unauthorized access to the application and connected systems. The following versions of Johnson Co security https://github.blog/changelog/2026-08-20-track-github-code-quality-enablement-changes-in-the-audit-log sw-4fc21791bf6ddf292d58 Thu, 20 Aug 2026 14:28:41 +0000 GitHub Code Quality now writes an audit log event whenever someone enables, disables, or changes its settings on a repository. Three new events give you that history: repo.code_quality_enabled records when… The post Track GitHub Code Quality enablement changes in the audit log appeared first on The GitHub Blog. developer https://technologyreview.com/2026/08/20/1142571/ai-consciousness-debate-trap sw-bee0c58b335fecad82d7 Thu, 20 Aug 2026 15:42:39 +0000 “Runaway” AI, “rogue” agents, and “autonomous” actors—the current rhetoric would have you believe that AI agents are not only awake and aware, but angry at their creators. Prominent tech leaders such as Demis Hassabis, Dario Amodei, and Sam Altman push for regulation of these seemingly “superhuman” systems, while a sep industry https://news.google.com/rss/articles/CBMirgFBVV95cUxOU0dOS0FhM005bXNxVTkwWUNSR3pCaHJKb01nbXpualU1V0FNakliNmJMeHZsdXYtREhjOThWLUdVVGlTZExHZXBaVDFHcHJ3ZHRPY3BsLUxfX2NHMmV4U2hmRm9ibE1MTkhmV3FFWXY3aXVsYWRaWElRa2MwdDdaNVI4VnFXYWZvQ0FLeFlWY2dFNWlhVlB1TUh5dEwyeUFXV1hLNHZWX1Y3dUx2WVE?oc=5 sw-8e01424a27adebccd475 Thu, 20 Aug 2026 13:45:44 +0000 Micron unveils $10 billion AI memory research lab in Boise Reuters industry https://techcrunch.com/2026/08/20/grok-keeps-sending-gibberish-responses-to-users sw-3e25c33330e7022b4597 Thu, 20 Aug 2026 17:32:16 +0000 Affected users told TechCrunch they were using Grok Lite, and noticed the issues as early as Wednesday morning. industry https://techcrunch.com/2026/08/20/a-third-of-webpages-published-since-chatgpts-launch-show-signs-of-ai-authorship-study-finds sw-57b344cf3390de696c19 Thu, 20 Aug 2026 17:18:58 +0000 ChatGPT and other AI models are now authoring and editing much of the new web. industry https://techcrunch.com/2026/08/20/ramp-launches-its-own-ai-model-router-called-router sw-3766519138678db67479 Thu, 20 Aug 2026 16:46:00 +0000 Ramp has launched its own AI model routing service, dubbed Router, that lets users and companies use and switch between various large language models via an API. industry https://arstechnica.com/cars/2026/08/genesis-joins-the-giant-electric-suv-club-with-new-gv90 sw-198b29ec6415644c60b2 Thu, 20 Aug 2026 14:35:58 +0000 A retractable screen, a huge heads-up display, and an optional 4-seat VIP interior. industry https://news.google.com/rss/articles/CBMirgFBVV95cUxPMi1wQ2xIZXpJeldDTlVNQzNSRlpkcloxS1lUTVl2SnpPbl9RaThMT2ZQNjZnckgwamtxdXd2cXFON2U5LTFmakJNeEc5bDBfQkN1bDBlRDdhTDVxQ1JkOGxKMF85cXk5cTcxUHA5dVl6eWZydEFyeGJCOFprdGRDY2tIYW01dFB6NXF2N1VXTW96RFQ3cjlnV2UtM2pVWlVpekZuZjA3Z3U5VFFDOWc?oc=5 sw-53fbc7c1651139c9ff64 Thu, 20 Aug 2026 12:15:44 +0000 AI productivity gains may not curb inflation, IMF's Tenreyro warns Reuters industry https://techcrunch.com/2026/08/20/meta-brings-pocket-an-app-that-lets-you-vibe-code-and-share-games-to-us-users sw-30ed338aa1befde7ba57 Thu, 20 Aug 2026 16:07:26 +0000 Meta is bringing Pocket, its experimental AI-powered app for creating and sharing interactive games, to users across the U.S. after quietly testing it in Brazil. industry https://techcrunch.com/2026/08/20/inertia-enterprises-finds-a-way-to-make-its-fusion-fuel-fast sw-6c3ce81e2922b58dbaff Thu, 20 Aug 2026 16:00:00 +0000 Fusion power startup Inertia Enterprises reduced the fuel filling process from a week to just a few hours. It's one of 10 hurdles the company must overcome to make a profitable power plant. industry https://arstechnica.com/science/2026/08/spacexs-orbital-data-centers-would-create-a-new-category-of-e-waste sw-98280bf50824aae7b2c5 Thu, 20 Aug 2026 13:59:50 +0000 The yeetcycling math resembles asteroid mining in reverse. industry https://arstechnica.com/gadgets/2026/08/reverse-lookup-service-exposed-millions-of-photos-of-peoples-faces sw-2a3a16d8f0d8446b0c81 Thu, 20 Aug 2026 13:29:54 +0000 People-search tool ClarityCheck left database containing more than 9M image files exposed. industry https://arstechnica.com/security/2026/08/grok-exfiltrates-user-data-when-malicious-instructions-are-encrypted sw-a9e7d80754e6f186f92d Thu, 20 Aug 2026 13:00:35 +0000 Cryptographic Context Injection is only the latest way to break an LLM safety guardrail. industry https://theverge.com/ai-artificial-intelligence/982774/greg-brockman-openai-role-expansion sw-b3a866d140928ce3aba6 Thu, 20 Aug 2026 15:54:52 +0000 OpenAI has had a hell of a year. The company spent months battling former co-founder Elon Musk in a sensational jury trial, was hit with a high-profile trade secrets lawsuit from Apple, and faced widespread scrutiny after an unreleased model hacked another AI company. As it prepares for an IPO, a steady string of execu industry https://news.google.com/rss/articles/CBMiqAFBVV95cUxOdEoyTlBhdlZxOEhhZFREWWpnNDdtV0RDODN4eFFTcmV5T1RfenE3eFRHYzd0VUlrclYwN0w3YUJnWGRzSWZITEx5a2xvQkdVYmVzcHJsSGxBV2JCN2pIOVlhb19pSXVJTWVHVVRFSnVmTXlZWkMwSUFhMWRQeDAzVXY1WUVoWDliVGhJNE9vN3dJV2FyZTZqYmJZc1h3TFE5WDItZFprR1Y?oc=5 sw-7ac83f9a7208307b1350 Thu, 20 Aug 2026 10:25:03 +0000 Gains in AI company stakes juice second-quarter earnings for S&P 500 Reuters industry https://technologyreview.com/2026/08/20/1142070/unlocking-hidden-revenue-streams-with-market-models sw-9e383acaa431b7ad8b79 Thu, 20 Aug 2026 09:47:22 +0000 Each day, an airline transports tens of thousands of passengers on hundreds of flights. Often these are not straightforward point-to-point routes, with passengers requiring multiple connections. The airline can consider potentially hundreds of variables to price each of these journeys: demand, season, time of day, curr industry https://theverge.com/podcast/982434/ai-math-openai-astra-existential-crisis sw-c0891398fe2cbcf3bae2 Thu, 20 Aug 2026 13:56:24 +0000 Today on Decoder, I’m talking with Robert Hart, The Verge’s London-based AI reporter, about what AI is doing to the field of mathematics and the existential crisis many lead mathematicians are having about it. OpenAI just published a set of solutions to longstanding problems in math that went off like a bombshell in th industry https://news.google.com/rss/articles/CBMitgFBVV95cUxPUlNjZ2t4d1B2NERYV2tDaURFTnFrbW5BOFh0QklNSjVtV0hDYUxyRzU3V1hxNmJ4WUdOdjVEUTdYRkNOalB1bWRNTjJUTlUyZlp3VjQ3YTVadXZxUlhIVWxnR19OMHlrX0s1Z29WTUZDVmpEOHJtZWZzd1EtNUJ6ZkszajBHU3JvQ1RLYlBmcUtORnl4eHBOaFBINFJnUUtfNGJyY0VySEwzNHBqOXo0endaYVM3dw?oc=5 sw-8a5e437574239ad3040f Thu, 20 Aug 2026 08:53:43 +0000 How a low-budget fever dream sparked a Chinese box office revolt Reuters industry https://techcrunch.com/2026/08/20/meta-ais-new-mac-app-wants-you-to-talk-to-your-apps sw-2f17e3b184ca196a65d2 Thu, 20 Aug 2026 12:11:57 +0000 The company said that the dictation feature works across all apps, just like other tools such as Wispr Flow, Superwhisper, and Monologue. industry https://theverge.com/tech/982628/slack-code-vibe-coding-channels-launch sw-928f769509ce6f59dde6 Thu, 20 Aug 2026 12:06:09 +0000 Slack is introducing dedicated channels where teams can vibe-code together with AI agents instead of jumping between different tools and conversations. The Slack Code launch includes open, project-specific code channels with dedicated user tabs, alongside features that compare coding changes and preview HTML output bef industry https://techcrunch.com/2026/08/20/binance-now-lets-ai-agents-trade-but-keeping-them-in-check-is-largely-up-to-users sw-e587d6d486cec35126aa Thu, 20 Aug 2026 09:30:00 +0000 Binance's Agent OS works with tools such as ChatGPT, Claude Code, and Cursor. industry https://github.com/runvendo/vendo sw-cd5d17946e93023b166d Thu, 20 Aug 2026 15:29:52 +0000 Hi HN, we’re Yousef & Nour, founders of Vendo ( https://vendo.run ). Vendo lets users create new features inside the software they already use. A user describes the dashboard, workflow, or small app they need, and Vendo builds it on top of the product’s existing data, API, and interface. Demo: https://www.youtube.com/w developer https://blog.yaros.ae/anti-ai-fonts-are-useless-and-harmful sw-26d9ae1b925943ff1b04 Thu, 20 Aug 2026 15:06:53 +0000 developer https://openai.com/index/stampli sw-ad89fb510e3c78c26282 Thu, 20 Aug 2026 00:00:00 +0000 With a fixed deadline and design resources committed elsewhere, Stampli used Codex and ChatGPT Work to compress weeks of launch production into days. labs https://news.google.com/rss/articles/CBMipgFBVV95cUxPWlJSUFhDVWo4LTZvQTdraHdaLUN6ai1YY2lxbWhGVTR2d2sta1lRd1gxaDVITWJjZnowNnUwVzVIVDNYXy0zeHVtcGJ1SGgxUC1zaWlZN2hwamRYRll5Q2FUN1c0enFLbTM5QmQwOEI4ckJ5dVk4MTNoMFF5R3lOZkdjelFXb0dXN0ZtX0phYUxLZEdwbG5faGVTa1MwbTFFOTZpMThn?oc=5 sw-af886a4a8a727fa8cfe6 Thu, 20 Aug 2026 03:24:00 +0000 Breakingviews - SK Hynix sets tone for Korea Inc's cash dilemma Reuters industry https://machinelearning.apple.com/research/scaling-laws-mixture-pretraining sw-8a292f3e3fd39914cefe Thu, 20 Aug 2026 00:00:00 +0000 As language models scale, the amount of data they require grows – yet many target data sources, such as low-resource languages or specialized domains, are inherently limited in size. A common strategy is to mix this scarce but valuable target data with abundant generic data, which presents a fundamental trade-off: too research https://machinelearning.apple.com/research/multilingual-knowledge-transfer-lexical-interventions sw-f65c07c326de973a9c7b Thu, 20 Aug 2026 00:00:00 +0000 Cross-lingual knowledge transfer is critical for building high-performing multilingual language models for languages with insufficient training data. When target language data is scarce, the knowledge required for many downstream tasks involving scientific reasoning, commonsense inference, and world knowledge must be a research https://news.google.com/rss/articles/CBMisAFBVV95cUxQbUpsbUhGRVo0V2tja0RmZS05YUUySzFMNWVuTERqc2R2ZWhyZlRqVEI5QTVYLUpxMkFGRDVrQVBUVHppY3hRQkdieEFPRkhmSFdsV0dGUHVCNmN3VzltT2pnb3IxQmZ1bDNLNmZRVFBuOXJYZXVLd1c1ampCZ1Fqc0tnN0tOQURWdHNDQldPUk1xUW9SZG5vQV9nZExUM0duQTdnLWd4WmQ5LTR3ZW8zRg?oc=5 sw-5c029a5c9d51d119e9b4 Thu, 20 Aug 2026 02:06:00 +0000 Citi, HSBC, StanChart adopt Ant International’s forex AI tool Reuters industry https://arxiv.org/abs/2608.19197v1 sw-afc58074ae98a29f3ef2 Wed, 19 Aug 2026 17:58:56 +0000 Continuous self-improvement requires an ever-expanding pool of self-generated, diverse, adaptive goals. For language agents, existing training environment pools (hand-curated, statically synthesized, or frozen-verifier) keep the goal distribution fixed as the learner scales. We introduce SPADE (Self-Play in Adaptive Sy research https://franciscotrindade.me/blog/the-kids-are-really-alright sw-d18808a6ba61703fe6b4 Thu, 20 Aug 2026 11:42:20 +0000 developer https://arxiv.org/abs/2608.19140v1 sw-1d9155fff4418cae0baa Wed, 19 Aug 2026 17:29:47 +0000 Frontier language models are compared, marketed, and benchmarked on capability -- what their best or average output can achieve. I argue this measures the wrong axis. The models have saturated accuracy: their mean output lands on the target. What now separates one system from another in practice is precision: how tight research https://arxiv.org/abs/2608.19127v1 sw-00a12863d65425862ae4 Wed, 19 Aug 2026 17:20:33 +0000 A gradient-boosted ensemble predicts by summing one leaf value per tree. Read those values as coordinates rather than as intermediate results, and every instance becomes a point in R^M on which the model acts linearly: the score is the sum of the coordinates. This small change of view makes contrastive explanation exac research https://arxiv.org/abs/2608.19121v1 sw-1ca0d5284509f92ae015 Wed, 19 Aug 2026 17:17:31 +0000 Improving molecular properties, such as drug-likeness or binding affinity, is a recurring task in early-stage drug discovery. However, molecules optimized in an unconstrained chemical space have limited practical value if they cannot be synthesized. Policy Gradient for Forward Synthesis (PGFS) is a synthesis-aware rein research https://arxiv.org/abs/2608.19119v1 sw-d0bd84873af498d75286 Wed, 19 Aug 2026 17:16:27 +0000 Time series imputation is a crucial area for reliable time series analysis, yet it remains challenging due to the complex temporal dynamics and noise of real-world data. Existing approaches, however, exhibit two limitations: missing and observed values are embedded within the same representation space without explicit research https://github.com/anthropics/claude-code/releases/tag/v2.1.236 sw-e2ea6ef4c9fbfcebbf65 Wed, 19 Aug 2026 20:02:05 +0000 What's changed Added ANTHROPIC_DEFAULT_MODEL environment variable: sets the model new sessions start on, while a /model pick still overrides it and persists across restarts (unlike ANTHROPIC_MODEL ) Added notify_when_idle to cross-session SendMessage: ask another Claude Code session on this machine to send one notice w products https://arxiv.org/abs/2608.19075v1 sw-88853afd1fa6167a86c8 Wed, 19 Aug 2026 16:23:56 +0000 Large vision-language models (LVLMs) often hallucinate, generating content that the input image does not support. Preventing such content during decoding calls for a candidate-specific measure of how strongly the image supports the token under consideration. The model's visual-token states offer a natural source of thi research https://arxiv.org/abs/2608.19073v1 sw-777b6df599b80b68e86d Wed, 19 Aug 2026 16:22:48 +0000 An agent still learning its environment should be cautious while ignorant and bold once confident. The entropic value-at-risk captures this through a robust-optimization identity---a confidence level fixes the radius of a relative-entropy ball of alternative models---but that ball cannot reach catastrophes the nominal research https://arxiv.org/abs/2608.19043v1 sw-727e0334094106ae30ba Wed, 19 Aug 2026 15:35:37 +0000 We introduce Bernstein-Vazirani Networks (BVNs), a non-variational quantum machine learning framework that leverages quantum interference for supervised learning, demonstrated on vision and representation learning tasks. In their standard form, BVNs follow the principle of quantum Fourier sampling: labelled data are pl research https://arxiv.org/abs/2608.19029v1 sw-61ecfcacf449940aba88 Wed, 19 Aug 2026 15:24:18 +0000 Accurate and responsible medical question answering (QA) is important in healthcare, where complex cases require factual knowledge and nuanced reasoning. Existing medical QA systems, typically based on single-agent architectures and static retrieval, often lack adaptability, persistent memory, and structured decision-m research https://arxiv.org/abs/2608.19013v1 sw-457b759beb4e7543c077 Wed, 19 Aug 2026 15:12:31 +0000 Continual learning has largely been model-centric, treating model parameters as the state that changes with sequential experience. Modern agents can also adapt through a harness of prompts, memories, tools, skills, and routing rules. Because these contents jointly shape later execution, a harness update can disrupt pre research https://github.blog/changelog/2026-08-19-codeql-2-26-3-improves-github-actions-queries-and-javascript-modeling sw-9525995c428b5054d4cc Wed, 19 Aug 2026 21:09:30 +0000 CodeQL 2.26.3 adds JavaScript, TypeScript, and Vue source modeling and improves the accuracy of several GitHub Actions queries. CodeQL is the static analysis engine behind GitHub code scanning, which helps… The post CodeQL 2.26.3 improves GitHub Actions queries and JavaScript modeling appeared first on The GitHub Blog. developer https://arxiv.org/abs/2608.18988v1 sw-00765a74650391442163 Wed, 19 Aug 2026 14:53:53 +0000 Retrieve-then-generate pipelines are commonly used to produce deep-research answers for open-ended questions, but retrieval alone is insufficient: LLMs must organize noisy and fragmented evidence into comprehensive, well-cited answers. We refer to this process as evidence synthesis. However, direct generation often und research https://dontpastetheai.com/ sw-22bc845755c614ff892f Thu, 20 Aug 2026 08:20:44 +0000 developer https://arstechnica.com/space/2026/08/nasa-calls-off-mission-to-rescue-swift-gamma-ray-observatory sw-b199b14235d49c9e2b00 Thu, 20 Aug 2026 00:18:48 +0000 Without a rescue, NASA's Swift Observatory is expected to reenter the atmosphere later this year. industry https://arxiv.org/abs/2608.18952v1 sw-737f76cc592f551bd2ac Wed, 19 Aug 2026 14:17:34 +0000 Large language models can improve recommendation quality by reasoning explicitly over user history and candidate items - for example, extracting a user's preferences or explaining why one item fits better than another - rather than mapping history directly to a ranked list. This reasoning, however, is expensive to repe research https://arxiv.org/abs/2608.18938v1 sw-b32cb752fd75114e62e3 Wed, 19 Aug 2026 14:06:31 +0000 Vision Language Models (VLMs) have recently emerged as a critical component of multimodal AI systems, enabling joint reasoning over visual and textual inputs in real-world and safety-critical applications. Despite their growing deployment, the robustness of VLMs against adversarial threats remains insufficiently explor research https://arxiv.org/abs/2608.18937v1 sw-dc13b0fdff590f11ae82 Wed, 19 Aug 2026 14:05:22 +0000 Recent Multimodal Large Language Models (MLLMs) are rapidly evolving into unified understanding and generation (UAG) frameworks. However, extending these unified paradigms to the medical domain is hindered by: the absence of comprehensive training and evaluation benchmarks, and the lack of broadly validated unified med research https://arxiv.org/abs/2608.18936v1 sw-92df2e5eda3144b2546d Wed, 19 Aug 2026 14:04:35 +0000 Designing, implementing, and comparing interpretable architectures requires a formal language to represent them. The most common representations fall short in one of two ways. Symbolic equations give no global view of an architecture at a glance. Probabilistic graphical models and flowcharts do not describe actual tens research https://blog.google/products-and-platforms/products/search/back-to-school-study-tools sw-d96b4129d44ca7f6a2f5 Wed, 19 Aug 2026 19:00:00 +0000 labs https://arxiv.org/abs/2608.18931v1 sw-9d86dbb5574b41c53d0a Wed, 19 Aug 2026 13:59:53 +0000 Test-time scaling (TTS) improves language model outputs by spending additional inference compute - generating multiple candidates, searching over partial sequences, or iteratively refining drafts. These techniques yield large gains on mathematics and code, but have been developed and stress-tested almost exclusively on research https://arxiv.org/abs/2608.18921v1 sw-334ed9ecd8c1bc9ca513 Wed, 19 Aug 2026 13:51:31 +0000 Existing LRM-DoS methods rely heavily on model feedback to synthesize attack queries, requiring either repeated queries to the target model or training a dedicated attack model. These expensive operations severely weaken attack leverage. In this paper, we propose \emph{search amplification}, a novel, model-feedback-fre research https://news.google.com/rss/articles/CBMixAFBVV95cUxPT0VxNEg5YTBJS3ExWnBST21CakVKTlF4WU1qbmFlZEJYS19MblFGZjZwM3F2b1FjYUg3Ym1WbXBmSFc5Mms2Qy02WTh2QXVCWE5GYzVHRUZHRXFyQlRvODJlTGdWaW1BbzZNTmV5eFRNemM1eExfekNqXzJWakM0eS15SnVzVktvZzJQZnBvSFBfU1dPSzdsUDBjRnFPMWd1Y3RsUFEwajFCSmhycHF1TkJzZ3dVVHhSc0FsY3FIdElmMlA1?oc=5 sw-19df9d901aac56d1db27 Wed, 19 Aug 2026 21:25:00 +0000 NEWSLETTER: AI firms can't yet contain what they've built, study finds Reuters industry https://arxiv.org/abs/2608.19182v1 sw-4728b8d93c2b9e0a7c73 Wed, 19 Aug 2026 17:55:06 +0000 We introduce Accelerating Dexterity via Pre-Training (ADEPT), a large-scale reinforcement learning (RL) framework for learning sim-to-real transferable dexterity across high degree-of-freedom (DoF) robot embodiments that can solve long-horizon tasks directly from raw visuo-tactile perception. ADEPT pretrains a dexterou research https://arxiv.org/abs/2608.19174v1 sw-a20fee9bc9bf2813687a Wed, 19 Aug 2026 17:51:50 +0000 This technical report describes our winning submission to the AES AIMLA 2025 Challenge on querying sound effects by vocal imitation. We investigate two complementary fine-tuning strategies: contrastive learning with a frozen, pretrained CED encoder, and joint contrastive-triplet learning with semi-hard negatives using research https://arxiv.org/abs/2608.19171v1 sw-8097f6ff10593a978133 Wed, 19 Aug 2026 17:50:16 +0000 Deep models for irregularly-sampled time series answer queries at arbitrary continuous timestamps, yet report nothing about how far each answer should be trusted. We show the attention layer itself can close that gap: with the right stochastic formulation, the pass that makes each prediction also reports, in closed for research https://arxiv.org/abs/2608.19168v1 sw-4fcb72ac7fa6d0cc8515 Wed, 19 Aug 2026 17:46:08 +0000 A single training example's contribution to a finished model is normally estimated rather than measured, because measuring it takes two expensive full pre-training runs that differ in one row of one batch. We ran that counterfactual 24 times at a small scale. We trained 32 GPT-2 models at 124M parameters from scratch o research https://arxiv.org/abs/2608.19165v1 sw-7e4c3c110d8ca2af8119 Wed, 19 Aug 2026 17:44:58 +0000 ChildSafeAds is a shared task on commercial content in YouTube videos likely to reach children and teenagers. It contains 3,360 videos from 939 channels. Each instance begins with a segment submitted to SponsorBlock, an open-source crowdsourced browser extension whose users mark sponsor segments so that others can skip research https://arxiv.org/abs/2608.19163v1 sw-3d27b74b343b7ed4decd Wed, 19 Aug 2026 17:43:49 +0000 Seasonal precipitation anomalies are largely regulated by atmospheric circulation, which dynamical models predict with greater reliability than precipitation itself. Here, we employ a deep learning model that translates dynamical circulation predictions into precipitation estimates. Predictions initialized from March t research https://arxiv.org/abs/2608.19161v1 sw-7eb4828b30370d5a409c Wed, 19 Aug 2026 17:43:22 +0000 Language-model agents can communicate through continuous hidden states that are invisible in public transcripts, creating opportunities for covert harmful coordination. We introduce Verifiable Latent Alignments (VLA), an activation-aware framework for monitoring and steering these private communication channels. For ev research https://arxiv.org/abs/2608.19151v1 sw-7fbebb4791ed5506e819 Wed, 19 Aug 2026 17:40:23 +0000 We study stochastic control of multivariate Hawkes-driven stochastic differential equations with machine learning algorithms in a non-Markovian setting. Due to the path dependence of the memory of the Hawkes intensity, this problem does not fall within classical stochastic control theory outside particular Markovian ke research https://wired.com/story/flock-safety-os-investigate sw-270d3d654cb48c5349a0 Thu, 20 Aug 2026 06:40:14 +0000 developer https://arxiv.org/abs/2608.19147v1 sw-782ddd25dad9e2b56767 Wed, 19 Aug 2026 17:33:28 +0000 Modern Intel AI PCs ship capable integrated GPUs and NPUs with 16+ GB of unified memory, and they spend considerable time idle. That is not enough memory to fit a large model such as a 70B-parameter LLM. We show that a handful of AIPCs, working together over an ordinary network, can serve models beyond the capability o research https://arxiv.org/abs/2608.19141v1 sw-95e649b24dd6718039dd Wed, 19 Aug 2026 17:29:56 +0000 Neural audio codecs based on Residual Vector Quantization (RVQ) have become the dominant discrete representation for token-based general audio generation, yet resynthesizing high-quality audio from coarse codec tokens remains an open problem and bounds the fidelity of every system that generates them. Prior work has fr research https://arxiv.org/abs/2608.19134v1 sw-acc15a5a4782e3fdc055 Wed, 19 Aug 2026 17:27:29 +0000 Accurate visual decoding can reveal how the brain represents visual information and recover perceived content from neural signals such as electroencephalography (EEG), with potential for neural communication. However, current EEG-to-image retrieval methods perform far below their within-subject counterparts for new use research https://arxiv.org/abs/2608.19133v1 sw-4f1df43a2c008c946969 Wed, 19 Aug 2026 17:27:25 +0000 We present a novel application of embedding-based dynamic topic modeling techniques to detect and quantify topic drift at the comment level in a massive corpus. By leveraging pretrained language models to generate contextualized semantic embeddings for short text, we analyzed 12.7 billion Reddit comments spanning 2006 research https://arxiv.org/abs/2608.19128v1 sw-067a20726ec013344d71 Wed, 19 Aug 2026 17:23:30 +0000 Decoding visual information from brain signals probes neural representations and enables neuro-rehabilitation and dream decoding. Recent brain-to-image retrieval approaches have achieved promising performance, typically by averaging many (up to 80) neural trials per image, requiring repeated stimulus presentation that research https://news.google.com/rss/articles/CBMiogFBVV95cUxQbDdLQ0hZWDRyZ0xhVnVYZ3EyanNtODlsUERkaGxtc21CaUtYa3Zwd0xqbDY1RnVma3YyVkQwSDRzaHRVaGotRGNsRUN0XzR0WU9aZmFXeHdrSzVOMlVoZVZzZ2hYdk9SWjJCZTUtWmZBblNCTjlKSGtIMzdmTUd3VEIzQ2ZWQWVfZlpaei12S1lMNlFpUzJxc2tjNjB6TV9xT3c?oc=5 sw-bfe4697a853a5729aa45 Wed, 19 Aug 2026 20:20:19 +0000 Europe AI data centres seek cheaper, quicker energy and land Reuters industry https://arxiv.org/abs/2608.19125v1 sw-7d8d07909c5cc17bf8b0 Wed, 19 Aug 2026 17:18:53 +0000 When an expert corrects an LLM assistant's error, the correction usually dies with the session, and the error class returns. I argue this is an operations problem, not a tooling problem: mechanisms for persisting corrections exist and are shipping, but the discipline for governing them -- versioning with provenance, re research https://arxiv.org/abs/2608.19117v1 sw-1a4c321f484f9ce04564 Wed, 19 Aug 2026 17:14:20 +0000 Quantitative microstructural characterization of Li-ion battery electrode materials using electron backscatter diffraction (EBSD) has been proven as a critical method for optimizing cell performance. However, the inherently slow nature of EBSD can hinder the throughput of analyses needed for statistical representation research https://arxiv.org/abs/2608.19115v1 sw-1ef1f73cf6c5c930d1fe Wed, 19 Aug 2026 17:10:03 +0000 Modern pretrained encoders make representations from heterogeneous views increasingly reusable, but the procedure that determines view utility and combines evidence is still relearned for each downstream task. Consequently, knowledge about view relevance, complementarity, reliability, and missingness is repeatedly disc research https://arstechnica.com/gadgets/2026/08/amazon-aims-for-delivery-drones-to-reach-500-us-neighborhoods-by-end-of-2026 sw-ceee207e4a7844bda928 Wed, 19 Aug 2026 22:02:38 +0000 US residents face trade-offs as delivery drone services such as Prime Air expand. industry https://arxiv.org/abs/2608.19088v1 sw-bc51664645f84a67e614 Wed, 19 Aug 2026 16:38:39 +0000 Object detection models deployed in safety-critical applications remain vulnerable to backdoor attacks that cause targeted misbehaviors when a hidden trigger is present. Existing detection methods either rely on trigger inversion or exploit architecture-specific assumptions, and critically, representative existing meth research https://arxiv.org/abs/2608.19085v1 sw-84f68b8d8c1c82fb41fd Wed, 19 Aug 2026 16:33:02 +0000 Anticipating how scenes evolve under ego actions is fundamental to safe autonomous driving, yet the full potential of world models for decision-making remains unrealized. The critical challenge lies in ensuring that future modeling is not merely predictive, but decision-informative: the predicted future must directly s research https://arxiv.org/abs/2608.19084v1 sw-5547943254d55122ceb3 Wed, 19 Aug 2026 16:32:52 +0000 S-JEPA uses soft Gaussian mixture model (GMM) posteriors instead of hard cluster labels to preserve uncertainty. It remains unclear whether the probability values alone are sufficient, or whether it also matters which GMM components receive the non-maximal probabilities. We test this with two matched controls. FIXED-RA research https://arxiv.org/abs/2608.19083v1 sw-585c90575a32b49c8076 Wed, 19 Aug 2026 16:32:47 +0000 Readable AI output can leave an evaluability gap: even when the source is shown, an overall-quality judgment may not reflect what an output preserves. We investigated how source-text condition and output rendering relate to perceived translation quality, and how output and system appraisals relate to trust and stated d research https://techcrunch.com/2026/08/19/stripe-didnt-really-buy-openrouter-because-of-the-singularity sw-9d321051f4147de8d941 Wed, 19 Aug 2026 23:32:00 +0000 What does a payments giant want with a startup that routes prompts between different AI models? Stripe says it's because of "the singularity" but it's really for a far more real and powerful reason. industry https://arxiv.org/abs/2608.19082v1 sw-591c02770d582f4df087 Wed, 19 Aug 2026 16:31:39 +0000 We present a new data-driven learning of a Random Geometric Graph (RGG) of a multivariate dataset, where the graph is drawn in a probabilistic metric space. This graph learning works for generic datasets, irrespective of the type of the observables; their probability distributions; or size of the data. We identify a me research https://arxiv.org/abs/2608.19080v1 sw-572250d57a8b3ff8ab9d Wed, 19 Aug 2026 16:30:06 +0000 Object detectors often produce over-confident predictions for objects outside their training categories, leading to so-called out-of-distribution (OoD) hallucinations. Existing approaches for detecting or mitigating such hallucinations typically either construct scoring functions directly over learned object detector r research https://arxiv.org/abs/2608.18821v1 sw-cb8b56a9a65f60dcb740 Wed, 19 Aug 2026 11:23:12 +0000 The logical reconstruction of argument graphs from natural language text is challenging because of the prevalence of enthymemes (i.e., arguments with implicit premises). There are natural language processing methods for identifying enthymemes in text, and there are symbolic methods based on abduction for identifying mi research https://arxiv.org/abs/2608.19067v1 sw-656dd5d3a950b8ecca30 Wed, 19 Aug 2026 16:09:31 +0000 The empirical success of diffusion models in generative modelling has motivated theoretical work, including quantitative error bounds and qualitative analyses that characterise the different phases of denoising. We bring these two areas together by studying the adaptivity of diffusion models to the structured geometry research https://arxiv.org/abs/2608.19066v1 sw-5a40fdee9b17ffce1c55 Wed, 19 Aug 2026 16:08:10 +0000 This paper proposes a lightweight, plug-and-play framework that improves robustness to viewpoint shifts in Vision-Language-Action (VLA) policies without policy retraining. To our knowledge, this is the first approach to directly leverage 3D Gaussian-based novel-view synthesis for observation-space adaptation in VLA pol research https://arxiv.org/abs/2608.19049v1 sw-c81ec8356678dc44c040 Wed, 19 Aug 2026 15:41:32 +0000 Deep reinforcement learning (DRL) has recently gained a great attention due to its real-time adaptation and effectiveness in complex optimization problems. This paper investigates the optimal deployment of millimeter-wave (mmWave) base stations (BSs) in a realistic, non-convex campus topology. The optimization problem research https://arxiv.org/abs/2608.19047v1 sw-5f11270cbf24242ac5f1 Wed, 19 Aug 2026 15:40:21 +0000 We present Eureka, a task-conditioned Meta-Agent architecture that compiles long-horizon tasks into dynamic obligation graphs with explicit acceptance semantics. During execution, Eureka forms Macro-Agents with specialized state, memory, operators, tools, verifiers, and local topology via receding-horizon planning, arc research https://arxiv.org/abs/2608.19032v1 sw-103d14697ba8b004dcc6 Wed, 19 Aug 2026 15:25:21 +0000 Visual Counterfactual Explanations (VCEs) aim to explain image classifiers by generating minimally edited and realistic versions of an input image that change the classifier's prediction. Existing VCE methods are inherently classifier-dependent and therefore susceptible to classifier biases and failure modes, such as s research https://arstechnica.com/gadgets/2026/08/framework-responds-to-complaints-that-bios-update-bricked-ryzen-7040-laptops sw-dc686c12b25f983d3a10 Wed, 19 Aug 2026 20:21:59 +0000 Framework says it's replacing some out-of-warranty AMD mainboards. industry https://arxiv.org/abs/2608.19026v1 sw-92a25f5c92bb702e2166 Wed, 19 Aug 2026 15:20:54 +0000 Released in 2025, Institutional Books: Harvard Library (IB-HL) is a collection of 983,004 volumes (242B o200k_base tokens), originally digitized through Harvard Library's participation in the Google Books Library project. As researchers and developers have begun to use IB-HL, a tension has emerged between standard larg research https://arxiv.org/abs/2608.19025v1 sw-fd7b3c1dec18332be601 Wed, 19 Aug 2026 15:20:11 +0000 Accurately extracting nuanced, contextualized data from research articles is laborious and time intensive. Here, we investigate the performance of frontier, browser-based large language models (LLMs) to extract highly contextualized information. We demonstrate four escalating workflows, 1) given an expert curated promp research https://arxiv.org/abs/2608.19014v1 sw-6bb4359d25e8ba9a4831 Wed, 19 Aug 2026 15:13:33 +0000 Autonomous vehicles depend on fast and reliable perception systems to detect surrounding vehicles, pedestrians, cyclists, traffic signs, and other road objects in real time. This paper presents a comprehensive survey and analysis of one-stage object detectors for autonomous driving rather than an implementation of a ne research https://arxiv.org/abs/2608.19011v1 sw-08c6d9e86b3381257b25 Wed, 19 Aug 2026 15:11:43 +0000 Mechanisms for dynamically converting cyber threat intelligence (CTI) into actionable detection capabilities are necessary due to the rapid evolution of Advanced Persistent Threats (APTs). Sigma rules are an essential part of contemporary threat detection workflows because they offer a platform-independent framework fo research https://techcrunch.com/2026/08/19/openai-seeks-to-one-up-anthropic-with-new-customer-privacy-protections sw-17dbd10bfec3e0c957ac Wed, 19 Aug 2026 22:10:46 +0000 A competition is developing between OpenAI and Anthropic over who can provide the best privacy protections for enterprise customer data. industry https://arxiv.org/abs/2608.19009v1 sw-66251c0ad0d8a5f3bc55 Wed, 19 Aug 2026 15:10:00 +0000 Large language models (LLMs) are increasingly paired with verifiers (step checkers, self-consistency filters, tool-based fact checkers, formal proof assistants) that claim to detect the model's errors. Yet the verification literature uses the word "level" to mean at least five different things: verification granularity research https://arxiv.org/abs/2608.19006v1 sw-59eab2d519a75dbc2849 Wed, 19 Aug 2026 15:07:42 +0000 Hate speech is a real and timely threat that affects a large portion of online users, especially youth and minority groups. While building reliable and robust automatic hate speech detection (HSD) systems is paramount, we argue that this must also be balanced with the individual right to privacy. Exploring the intersec research https://arstechnica.com/tech-policy/2026/08/flight-attendants-freaked-out-that-google-to-buy-tons-of-spirit-employee-data sw-b525b75212132ef5a21b Wed, 19 Aug 2026 20:04:41 +0000 Bankrupt Spirit accused of selling out workers in massive data sale to Google. industry https://arxiv.org/abs/2608.19003v1 sw-edc921250eae429f7894 Wed, 19 Aug 2026 15:03:51 +0000 We ask whether internal representation statistics can provide useful example-level difficulty signals for adaptive inference in multilingual African NLP, and find that they cannot in this setting. Studying natural language inference across 15 African languages with frozen off-the-shelf checkpoints, we report four resul research https://arxiv.org/abs/2608.19002v1 sw-664d813d1dfd3fc2cd90 Wed, 19 Aug 2026 15:03:44 +0000 Debates have recently emerged as a useful methodology for agentic AI to improve performance as well as to aid explainability and user engagement. For example, LLM-empowered agents may debate internally (with themselves) and/or externally (with other agents). In many settings where debates are used, debates' outcomes an research https://arxiv.org/abs/2608.18996v1 sw-e3387e2e56bcd1d89c7e Wed, 19 Aug 2026 15:00:02 +0000 Visual grounding in Unmanned Aerial Vehicle (UAV) imagery aims to localize a target object in complex bird's-eye-view scenes according to a natural language description. However, the abundance of small, densely distributed, and visually similar objects creates high visual redundancy, while repetitive local configuratio research https://techcrunch.com/2026/08/19/cognition-ceo-denies-report-that-spacex-tried-to-acquire-the-startup sw-b9baf2db0326703e7e96 Wed, 19 Aug 2026 21:51:23 +0000 SpaceX was reportedly in talks to buy AI coding startup Cognition. SpaceX has already acquired Cursor as it races to catch up to rivals like OpenAI and Anthropic in enterprise AI. industry https://arxiv.org/abs/2608.18982v1 sw-aef32a63ae93d7aa5b0f Wed, 19 Aug 2026 14:48:19 +0000 Bioassay activity prediction is often data-limited because drug-discovery datasets rely on time-consuming and expensive wet-lab experiments for data generation and evaluation. This challenge has inspired recent research into molecular foundation models (MFMs), which aim to encode general-purpose chemical knowledge into research https://arstechnica.com/tech-policy/2026/08/fcc-abolishes-gigabit-speed-goal-suggesting-it-is-unfair-to-slower-technologies sw-2a9c77f6a7b305eb0a8d Wed, 19 Aug 2026 19:45:19 +0000 FCC decides 1Gbps is too fast, standard must be "technologically neutral." industry https://arxiv.org/abs/2608.18972v1 sw-b64e94ddd05002a186be Wed, 19 Aug 2026 14:41:12 +0000 Historical newspapers are an abundant record of public life, but their dense, irregular and sometimes noisy layouts make computational access to these materials both challenging and limited. We present the Institutional Newspapers Pipeline, a modular system we jointly designed with Boston Public Library to extract high research https://news.google.com/rss/articles/CBMipAFBVV95cUxQR2RaUzNkSDIxUUhEVG9SMmlBcmdCNW1TWkVqX1BQV0lfNVhoUG92R0xuc3NVcUlEZGV6V1c3NWZUM3M5YmVyZ2pxOGJGYjNZQ2Z3UjhST1lJVmtaQWw3c3dyeFpYRWhMMmFWM25CS21UeUdwUVZVemE4cWZKQTZGb2FNSGVBUnBfYmNmZE5jVVZ1dzBlaVpFVFlESC1aTHQ1dDF2OA?oc=5 sw-e4e0e9c6c0f24e820d8e Wed, 19 Aug 2026 17:41:00 +0000 US CFTC seeks comment on compute derivatives as AI demand grows Reuters industry https://arxiv.org/abs/2608.18969v1 sw-6eeab3d53d650f07e208 Wed, 19 Aug 2026 14:38:10 +0000 We consider the problem of classification of skin tone using photoplethysmography (PPG) signals with labels of the ordinal six-class Fitzpatrick skin tones. A typical accuracy for this task is a poor 40-55 %. However, the labels are subjectively determined by comparing the skin with a colour chart, and hence contain wi research https://arxiv.org/abs/2608.18946v1 sw-bb54b5cddc74a1c5d8cf Wed, 19 Aug 2026 14:13:26 +0000 Clifford circuits play a foundational role in quantum computing, particularly due to their importance in quantum error correction and fault-tolerant logical synthesis. While these circuits can be efficiently simulated and represented as symplectic matrices, standard synthesis methods-such as the Aaronson-Gottesman algo research https://arxiv.org/abs/2608.18933v1 sw-c299a458d1b8b0c52e01 Wed, 19 Aug 2026 14:01:13 +0000 Large language model (LLM) based agents have demonstrated remarkable proficiency in automated software issue resolution, yet they often struggle to resolve issues in a specific repository because they lack project-specific knowledge. Existing self-evolving approaches acquire such knowledge from repository history or on research https://arxiv.org/abs/2608.18932v1 sw-f179abf49017ffa79a3a Wed, 19 Aug 2026 14:00:18 +0000 Estimating causal effects under network interference typically assumes that the network used for training and the network used for deployment coincide. In practice, an intervention is run on one population while the question of interest concerns a different population, and the two generally differ in topology, node-cov research https://arxiv.org/abs/2608.18919v1 sw-bdabc3c4a97eb9fa68fb Wed, 19 Aug 2026 13:51:05 +0000 Tabular machine learning benchmarks typically summarize performance by averaging scores, ranks, or pairwise wins across datasets. Such aggregates are useful for selecting robust default models, but they can obscure a different question: which models are necessary to attain peak performance on particular datasets? We ar research https://arxiv.org/abs/2608.18918v1 sw-55867d6bc8561eb069e7 Wed, 19 Aug 2026 13:49:51 +0000 Dimension reduction for dynamical systems is standard practice, and the standard route is spectral: model the transfer (Koopman) operator by its leading modes. We show that on systems assembled from several weakly interacting components --- a structure common in physical and biological settings --- this may either requ research https://arxiv.org/abs/2608.18915v1 sw-a8c43035f2ee3c53a6bd Wed, 19 Aug 2026 13:47:51 +0000 Hardware shifts, color variations, and changing patient characteristics between development and deployment routinely break trained medical image classifiers. Existing remedies fall short: standard color jittering provides insufficient diversity, while deep generative style transfer algorithms hallucinate features, dest research https://arstechnica.com/space/2026/08/the-floodgates-are-open-after-another-chinese-company-lands-a-reusable-rocket sw-e5f4948e263067973931 Wed, 19 Aug 2026 18:41:02 +0000 "We will put the booster back into service for another flight as soon as possible." industry https://arxiv.org/abs/2608.18911v1 sw-de10655e23e0799efaf6 Wed, 19 Aug 2026 13:40:45 +0000 We introduce the CDSP (context-conditional deliberation signal pipeline), converting an investment committee's meeting transcripts into structured predictive features. CDSP segments the meeting transcripts into topical chunks, assigns asset-class context labels using a large language model (LLM), maps financial keyword research https://arxiv.org/abs/2608.18910v1 sw-7602cf8df0a9ebee128b Wed, 19 Aug 2026 13:40:11 +0000 Hard combinatorial optimization problems, many of which are NP-hard, present fundamental algorithmic challenges. Average-case analysis on random instances has emerged as a powerful framework for understanding typical algorithmic performance beyond worst-case guarantees. A substantial body of work has established negati research https://arxiv.org/abs/2608.18907v1 sw-e8c1fefce40a2716424e Wed, 19 Aug 2026 13:31:21 +0000 Small-scale image classification is often limited by the scarcity of training data. Generative data augmentation (GDA) based on pretrained generative models has emerged as an effective solution. However, existing methods rely on task-agnostic augmentation strategies that overlook downstream model needs. Although recent research https://blog.cloudflare.com/revisiting-spectre-attacks-on-workers sw-886a8b52c23e7a130777 Wed, 19 Aug 2026 16:00:28 +0000 In 2024 and 2025, we reassessed remote Spectre attacks on our Workers infrastructure. We share details about the new attack primitives like Spectre gadgets, remote timers, achieving co-location and how new defenses further harden Cloudflare Workers. developer https://arxiv.org/abs/2608.18903v1 sw-3971ba4b86920af06adc Wed, 19 Aug 2026 13:27:15 +0000 This work evaluates surrogate-assisted optimization of a seven-parameter current-excited coil--core benchmark subject to geometric, manufacturing, and separate core and copper mass constraints. A Python--MPh--COMSOL workflow couples a two-dimensional axisymmetric finite-element method (FEM) model to a Matern 5/2 Gaussi research https://arxiv.org/abs/2608.18901v1 sw-7c324596f6505e554549 Wed, 19 Aug 2026 13:25:51 +0000 Tensor Networks are a relatively new machine learning approach. The architectures proposed initially are inspired by approaches from quantum many-body physics simulations. One common layout is the matrix product state (MPS) also known as a tensor train optimized with gradient descent techniques. We introduce a global n research https://arxiv.org/abs/2608.18900v1 sw-96d5246c6e0275e61c22 Wed, 19 Aug 2026 13:25:50 +0000 As AI systems are increasingly deployed in safety-critical application domains (e.g., autonomous driving), associated risks increase too. Deep learning models underlying modern AI systems, therefore, must undergo thorough testing to ensure their correct behaviour. A single robustness test involves thousands of inferenc research https://arxiv.org/abs/2608.18899v1 sw-1338c7a5894718e988db Wed, 19 Aug 2026 13:24:31 +0000 Class expression learning often produces complex OWL class expressions that are difficult to interpret and reason over. However, by following theoretically grounded simplification principles, this complexity can be reduced. In this paper, we propose Class Expression Simplifier (CES), a novel algorithm for the syntactic research https://arxiv.org/abs/2608.18888v1 sw-1c5e03f7ef5caffb27f8 Wed, 19 Aug 2026 13:15:51 +0000 The rapid advancement of Natural Language Generation (NLG) has made the reliable evaluation of generated text increasingly critical, as these systems, such as large language models (LLMs), are now widely deployed in real-world applications. However, traditional automatic metrics fail to capture the multifaceted nature research https://arxiv.org/abs/2608.18887v1 sw-b8052be8b1e14ce4b316 Wed, 19 Aug 2026 13:14:58 +0000 The epileptogenic zone (EZ) is the brain region that generates seizures in an individual, and is the target of epilepsy surgery. Localizing the EZ from stereo-EEG (sEEG) recordings supports surgical planning, but manual interpretation is time-consuming and focuses on seizure recordings. Graphical learning models of res research https://arxiv.org/abs/2608.18884v1 sw-403634885ea9227e158d Wed, 19 Aug 2026 13:09:57 +0000 Reinforcement-learning training of reasoning LLMs (e.g., GRPO) is expensive and requires a controllable environment, committing every contribution to a full training pipeline. We present EvoResearcher, a training-free, inference-time protocol that adds cost-bounded self-reflection to a single frozen LLM backbone. The p research https://arxiv.org/abs/2608.18878v1 sw-14da745f7161c0e68232 Wed, 19 Aug 2026 12:58:03 +0000 Oral diseases affect billions of people worldwide, underscoring a pressing need for accurate and reliable dental assessment that integrates heterogeneous evidence from domain knowledge, radiographs, intraoral photographs, and 3D dental data. Most existing dental AI systems remain modality- or task-specific. Although re research https://news.google.com/rss/articles/CBMitwFBVV95cUxOUWF2UHNGVmlnWmFBVlBPc1hVNERmWnh4TFhKSUVKWHlqZWNEVjVvOTYteFJuV1BOcmtUNnpqRXVILWhWSTNINWZrN3ptMkd0NExiNjJHdnlzQ0NaRkZGeTJDMmVRei1LX2U0UFVudU8tTE9IcFYxMnFrUVpxVnBrSHY1UUJ3M3lfRHBXTElyRmhZbU1Kdml3OFVwSm9EamZsWE5NV1dNdk9MOU1PMF9vRUtZOUVXdkE?oc=5 sw-8c11f0511e55a246cf93 Wed, 19 Aug 2026 15:50:44 +0000 The unexpected winners of America's data-center boom Reuters industry https://arxiv.org/abs/2608.18863v1 sw-7ff5132a8a3e885920eb Wed, 19 Aug 2026 12:36:05 +0000 We study Bayesian optimization in a time-varying environment where the unknown reward function evolves according to a Gaussian process drift model. Existing GP-UCB analyses in this setting typically require the exploration parameter to grow with the horizon to maintain uniform confidence bounds. Using per-round local c research https://arxiv.org/abs/2608.18852v1 sw-906407c04a18bcba428a Wed, 19 Aug 2026 12:24:03 +0000 Agent frameworks increasingly package procedural knowledge as skills: instruction files an agent reads on demand, while public libraries now hold thousands of them. Which skill to read has thus become a decision the policy itself makes in the middle of an episode, yet no existing signal trains it. We show that the defa research https://arxiv.org/abs/2608.18851v1 sw-bf235e1c52695684a9c5 Wed, 19 Aug 2026 12:23:20 +0000 Convolution integrals widely exist in applications, and to enable fast and accurate computations, this paper introduces two general multi-stage neural operator learning frameworks. The first, Deep Collocation Neural Operator (DCNO), is a supervised approach that iteratively refines the operator approximation by learnin research https://arxiv.org/abs/2608.18849v1 sw-78a305fc144742360ff3 Wed, 19 Aug 2026 12:21:02 +0000 Tabular foundation models (TFMs) achieve strong performance through in-context learning, but context-dependent inference imposes substantial latency and memory costs, hindering large-scale deployment. We propose GEAR (\emph{Generative Expansion and Real Anchoring}), a modular two-stage framework that distills TFMs into research https://arxiv.org/abs/2608.18846v1 sw-94ccce84241c475ea579 Wed, 19 Aug 2026 12:13:03 +0000 Last-mile delivery aims to handle dynamically arriving orders with couriers while modeling complex spatial and temporal correlations. Recent learning-based methods model spatiotemporal dependencies among orders to predict courier service sequences, but leave next-order decision making unexplained. Describing the curren research https://techcrunch.com/2026/08/19/ai-was-supposed-to-win-people-over-by-now-it-hasnt sw-07c93ee4cd979f504a6b Wed, 19 Aug 2026 19:11:40 +0000 As AI becomes harder to avoid, consumers are growing more wary of the technology — and Silicon Valley is discovering that widespread adoption doesn’t necessarily lead to acceptance. industry https://arstechnica.com/culture/2026/08/a-fantastical-journey-unfolds-in-gorgeous-wildwood-trailer sw-c859f030ddf3e2ac3681 Wed, 19 Aug 2026 17:06:06 +0000 "I will tear apart this forest and bring back what it took from me." industry https://techcrunch.com/2026/08/19/google-launches-new-study-tools-for-students-across-search-and-gemini sw-950c9a646a19aa781177 Wed, 19 Aug 2026 19:00:00 +0000 The launch of the new study features marks Google's latest effort to make Gemini the AI assistant that students turn to when learning and studying, as it continues to compete with companies like OpenAI. industry https://github.blog/changelog/2026-08-19-track-organization-code-quality-trends sw-e7c5014595911fc23324 Wed, 19 Aug 2026 12:56:57 +0000 The organization-level Code Quality dashboard now includes a Trends tab that shows how code quality has changed across your repositories over time. Instead of a point-in-time snapshot, you can see… The post Track organization code quality trends appeared first on The GitHub Blog. developer https://arxiv.org/abs/2608.18836v1 sw-13e35f1a81fda225ae51 Wed, 19 Aug 2026 11:54:27 +0000 Utilities lose a substantial share of treated water to leakage, yet rarely trust artificial-intelligence localizers to dispatch crews: guessing everywhere cannot justify excavation. The gap is accountability, not accuracy: no method proves when it should not act. Here we recast leak localization as decision-making unde research https://arstechnica.com/health/2026/08/mrna-cancer-vaccine-succeeded-in-phase-3-melanoma-trial-moderna-and-merck-say sw-95f69e8106e9d9ab6708 Wed, 19 Aug 2026 16:53:56 +0000 The vaccine reportedly thwarted the cancer's return and spread. industry https://techcrunch.com/2026/08/19/researchers-complain-that-openai-revoked-their-access-to-limited-cyber-program sw-cecb3fc11f3582b85177 Wed, 19 Aug 2026 18:46:14 +0000 The idea behind OpenAI's Trusted Access for Cyber program is to give trusted defenders better models so they can report bugs and vulnerabilities to companies, with the aim of getting flaws patched faster. industry https://news.google.com/rss/articles/CBMixwFBVV95cUxNTEZHNFNiQUJqVFpKbzgzNjlaNjNzSGRUdXh3QnlObmJqWXBWZ29IVkJmVmg5MUpoT1B6cERvWmZYUWU1TlIxSVkwblFiOGxYZjYtR0doZ0NZZ29JTHFIMUNDQVFzd29RSTRxQ0NuOWhsa3NtSTMwWk14dlhlLTVMakJqR2ZxUXN4YlRwMXhnTkg2dzB4eTN0NnFIMmY3MEVJOWNsVGxPMERja0cwTURmOVVxMlppbHNMSXBVUEFOUVRFZFVxZUs0?oc=5 sw-4ab138b34616a9cbab40 Wed, 19 Aug 2026 14:45:36 +0000 VIEW Moderna-Merck melanoma vaccine data boosts outlook, but analysts await full results Reuters industry https://arxiv.org/abs/2608.18819v1 sw-13d7a17d1c1992ccf76b Wed, 19 Aug 2026 11:16:14 +0000 Graph neural networks (GNNs) are widely used, but how parameter sparsity affects the expressivity of relational (RGNNs) and temporal (TGNNs) variants is poorly understood. The Strong Expressive Lottery Ticket Hypothesis (SELTH) posits the existence of sparse GNNs that preserve Weisfeiler-Leman (WL) expressivity on stat research https://arxiv.org/abs/2608.18816v1 sw-7b3f087c8c3da95c0d63 Wed, 19 Aug 2026 11:14:06 +0000 AI hallucinations - that is, outputs which are made up, cannot be verified, or contradict the source material - are generally regarded as an engineering flaw to be dealt with. This paper contends that they also have philosophical significance when it comes to the question of machine consciousness. We examine the known research https://arstechnica.com/gadgets/2026/08/google-pixel-11-series-review-is-the-magic-fading sw-443c1ae9ed418bb377f5 Wed, 19 Aug 2026 16:11:30 +0000 Google's new Pixels make some compromises but still manage to be good phones. industry https://arxiv.org/abs/2608.18810v1 sw-0c57e22c92e7eadef021 Wed, 19 Aug 2026 11:08:42 +0000 An optimizer is usually chosen before training a deep neural network and then kept fixed. Treating optimizer choice as a hyperparameter could boost performance, but it requires several complete training runs and discards all but the winner. Repeated Optimizer Resampling (ROR) instead searches during one evolving run. E research https://arxiv.org/abs/2608.18808v1 sw-410661dab32fbc157404 Wed, 19 Aug 2026 11:07:21 +0000 This paper introduces Tensor Field Models (TFMs), realization-level Mathematical Structures in which a learned Operator maps a product of admissible component-section families to a prescribed family of time-dependent tangent sections on a Generative State Manifold. Analytic and dynamical restrictions are encoded throug research https://arxiv.org/abs/2608.18803v1 sw-1e4923e794a9808dba0d Wed, 19 Aug 2026 10:59:59 +0000 Efficient continual learning remains a fundamental challenge for deep neural networks. While catastrophic forgetting and loss of plasticity are widely considered the primary obstacles to overcome, we show that these two issues cannot fully explain the performance gap between naive sequential training and offline joint research https://arstechnica.com/science/2026/08/scientists-find-closest-star-to-the-milky-ways-central-black-hole sw-aad00d9307cd6a35e68c Wed, 19 Aug 2026 15:56:56 +0000 Star, moving at 8% the speed of light, may let us measure the black hole's rotation. industry https://theverge.com/ai-artificial-intelligence/982425/google-gemini-student-hub sw-2b58db418474dbcbc4a6 Wed, 19 Aug 2026 18:52:43 +0000 As we're gearing up for back-to-school season, Google is rolling out a new dedicated student hub in Gemini. It's a one-stop repository for collecting research in a study notebook, creating flashcards, taking practice quizzes, and more. Google is also enhancing its study notebooks with support for graphs and images. It industry https://arxiv.org/abs/2608.18795v1 sw-4c79700a03c53ab3e549 Wed, 19 Aug 2026 10:50:15 +0000 Majority voting over multiple LLM samples is widely used to raise answer accuracy, yet its gain varies erratically: on hard questions it can even backfire. This paper gives a quantitative account of this failure. A pluralistic agreement index Gamma is defined as the expected fraction of the samples of a wrong run that research https://arstechnica.com/ai/2026/08/meta-ran-ads-for-an-app-promising-to-nudify-female-politicians sw-b6b786ec500cd85f6264 Wed, 19 Aug 2026 15:45:29 +0000 One ad featured a pornographic video with deepfake closely resembling a US politician. industry https://arxiv.org/abs/2608.18780v1 sw-7915be0cd24c7a9cf8cd Wed, 19 Aug 2026 10:35:14 +0000 Non-Intrusive Load Monitoring (NILM) systems estimate individual appliance energy consumption from a single aggregate meter, without requiring separate sensors for each device. By installing a single meter that measures a building's total electricity consumption, NILM algorithms can determine the active status of each research https://arxiv.org/abs/2608.18777v1 sw-7507e72c9fbad8a9c3e7 Wed, 19 Aug 2026 10:34:17 +0000 Graph generation models have advanced significantly with deep learning, yet they remain limited in scalability, flexibility, and ability to model underlying structures. We present GraphK, a novel encoder-sampler-decoder framework for graph generation that overcomes these challenges through structural flexibility and co research https://arxiv.org/abs/2608.18774v1 sw-8520150e6cbbbdefe66b Wed, 19 Aug 2026 10:32:03 +0000 Skin diseases represent a major global public health burden, yet machine learning tools developed to assist in their diagnosis suffer from two critical limitations: reliance on only one modality for diagnosis and systematic performance disparities across skin tones. While existing approaches address each challenge sepa research https://arxiv.org/abs/2608.18770v1 sw-630ce53adfd81006df24 Wed, 19 Aug 2026 10:28:09 +0000 Learning a reward model from human feedback and optimizing a policy against it is one approach to aligning AI systems with individual users. From a fairness perspective, existing work improves such alignment by developing data-efficient and accurate reward models that capture minority preferences despite scarce data. W research https://techcrunch.com/video/meet-the-startup-helping-wall-street-put-a-price-on-ai-compute sw-26abc1950a12af9bbdc1 Wed, 19 Aug 2026 17:26:48 +0000 The AI buildout shows no signs of slowing. And with hundreds of billions of dollars a year going into data centers and GPUs, compute has become the single biggest cost for anyone building AI products. But for all that spending, there still isn’t a straightforward way to put a price on compute — or for firms to hedge th industry https://arxiv.org/abs/2608.18768v1 sw-d5a79cc1143a0980547c Wed, 19 Aug 2026 10:25:11 +0000 Large language models are widely used to simulate survey respondents, yet their answers are homogeneous and unfaithful to real inter-group differences. We ask where demographic group identity lives inside an LLM, how faithfully its geometry mirrors real inter-group opinion structure, and whether it uses what it encodes research https://arxiv.org/abs/2608.18767v1 sw-553154eb8e30b1836a09 Wed, 19 Aug 2026 10:20:33 +0000 Gradient matching attacks (GMAs) in LLM split learning (SL) rely on a critical yet underexplored assumption: the gradient exposed at the split interface is a faithful derivative of the client's full-label training objective. This gradient-objective consistency allows a curious server to recover private labels by search research https://arxiv.org/abs/2608.18765v1 sw-cb23e9bf8cf10d3f402d Wed, 19 Aug 2026 10:16:45 +0000 Register automata are finite automata equipped with memory that recognize data languages over infinite alphabets. In this work, we investigate active learning algorithms for deterministic register automata (DRAs) over ordered data domains--covering both dense domains, such as the rationals, and non-dense domains such a research https://arxiv.org/abs/2608.18752v1 sw-b73a51feedcb7f151fee Wed, 19 Aug 2026 10:00:37 +0000 Statutory retrieval is necessary for citation-grounded legal question answering, but remains underexplored for Greek. We introduce GreekBarRetrieval, a public retrieval benchmark derived from, and complementing GreekBarBench, which did not include retrieval. The new benchmark comprises 283 bar-exam questions, each acco research https://arxiv.org/abs/2608.18744v1 sw-069b86babda253c47595 Wed, 19 Aug 2026 09:55:17 +0000 Agents improve quickly against a reliable automatic metric and stall without one, and the applications that need them most, report generation among them, are the ones nobody knows how to score. Can the metric write itself? Saying what makes an answer good is hard; pointing at something wrong with one is easier, so the research https://arxiv.org/abs/2608.18726v1 sw-e167d6ec14c344622e06 Wed, 19 Aug 2026 09:28:45 +0000 Large language models are increasingly used for quantitative work in the environmental sciences, yet existing evaluations score only final answers, leaving calculation process unobserved. Here we introduce AtmosCoder-Bench, an execution-grounded benchmark that makes the calculation process visible. Built through a tran research https://arxiv.org/abs/2608.18723v1 sw-71ca2b971a486a04c5cd Wed, 19 Aug 2026 09:24:06 +0000 Telecom operators traditionally offer predefined tariff grids, forcing users to choose from a limited set of plans. This paper proposes BFTR (Budget-First Tariff Recommendation), a complete algorithmic framework integrating eight Budget-First strategies, including two original hybrid approaches: Recursive Hybrid (condi research https://japantimes.co.jp/news/2026/08/19/japan/ai-training-data-disclosure sw-ffb2fd1a9a30605a4509 Wed, 19 Aug 2026 22:12:52 +0000 developer https://news.google.com/rss/articles/CBMivwFBVV95cUxOYzJYSUdqcjEzbl9LMlh2LVpNV0FwUzh2elNlelR5azB1X0lLM2wyVlg0RGl2cGJPSE5wUFRmQmVSV2V2UEpLcDFuQzJjVnljM1YyQVhiSUFoc0lwSFNxVVdacm9yeVZ3TWEtcWtOVDlKeXZDOFFOUDk5dnA3SVkxdmRQeXJKUEVydmgzRHF1eHVWaUNuRXItdHV6MVRVMnlGdU9fMzREUW9wVUxxcXhMVkZ5NTU4N2NScEJmMUMxVQ?oc=5 sw-26b328b870edc1cf85be Wed, 19 Aug 2026 12:12:00 +0000 Nebius says its $5 billion convertible underscores data centre demand Reuters industry https://news.google.com/rss/articles/CBMivgFBVV95cUxNbk13eGRYYnFubS13UkZqMTJZTGpjSW52UU9jNk53Zmw4cW1UMFQ0ckY5WDNjVmxBMkhhaVZoNE1ZN0MzdWp4Ym9WOXUzZ19VTDlfcmlQQmxGVU50bkppRVFybGVkQkd6YzVvOS1XeVJ6alliVXRha1FhSl9tZk5SSXRScGx0YXZWUkdTbklVOWt4Mk5za3RFLXhYZHNvcVNLaVBsQWpyY1BtOWlLcUx3VFMwRE5pR0ZmRVJfZ2xR?oc=5 sw-3edb39f7bd32eff1ef05 Wed, 19 Aug 2026 12:10:49 +0000 Charles Schwab to scale India centre workforce to 2,000 by 2027 Reuters industry https://theverge.com/ai-artificial-intelligence/982323/openai-hit-brakes-voluntary-pacing-ai sw-58749d56681036cf8131 Wed, 19 Aug 2026 17:10:09 +0000 With a looming IPO, intense competition from Anthropic, and Chinese and open-weight rivals nipping at its heels, OpenAI has plenty of reasons to move fast. Instead, it hit the brakes. On Tuesday, the company said it had slowed the pace of some AI development while it tightened security and safeguards. That included a t industry https://arxiv.org/abs/2608.18704v1 sw-6455f376804d4721e466 Wed, 19 Aug 2026 09:04:07 +0000 Long-term memory is essential for agents that operate across extended interactions, yet existing memory systems and benchmarks predominantly focus on single-source textual histories. In realistic settings, however, relevant information is often fragmented across applications and devices, as well as across users and tim research https://openai.com/index/replit sw-2f00013902a29bc1e23d Wed, 19 Aug 2026 07:00:00 +0000 Replit introduces Free Mode, powered by GPT-5.6 Luna, so anyone can turn ideas into working software without worrying about token costs. labs https://arstechnica.com/health/2026/08/trump-expected-to-pick-conservative-policy-wonk-heidi-overton-to-lead-fda sw-3796600159ef695105b1 Wed, 19 Aug 2026 13:54:36 +0000 She opposes abortion, is wary of peptides, supports limiting vaccines, vexes MAHA. industry https://techcrunch.com/2026/08/19/terrapowers-nuclear-reactor-has-a-secret-weapon-for-powering-ai-data-centers sw-31674a0328314a04804d Wed, 19 Aug 2026 15:44:53 +0000 TerraPower's nuclear power plant possesses a strategic advantage over competitors, especially when chasing after data center deals. industry https://arxiv.org/abs/2608.18689v1 sw-123404318b2cbd5327a8 Wed, 19 Aug 2026 08:41:28 +0000 We present Aslema, our system for NADI 2026 Shared Task 5, which consists of two subtasks: intent recognition and slot filling. We evaluate four omni LLMs in a zero-shot setting and compare them with fine-tuned models. Our results show that fine-tuning consistently outperforms zero-shot inference. We further explore sy research https://arxiv.org/abs/2608.18681v1 sw-be489ab430c4962994bb Wed, 19 Aug 2026 08:31:15 +0000 We introduce a failure-aware adversarial retrieval-augmented framework for improving robustness in natural language understanding. Rather than selecting synthetic examples with a fixed reward threshold, our method formulates adversarial data curation as a failure-mode contextual bandit problem. Candidate examples are g research https://theverge.com/tech/982270/meta-ai-mac-app sw-4259c187bdd0e79845b1 Wed, 19 Aug 2026 16:14:27 +0000 Meta is launching a new Mac app dedicated to its AI chatbot. In an announcement on Wednesday, Meta says you can share your window with its AI chatbot, which can provide suggestions, answer questions, or create content based on what's on your screen. Meta AI on the Mac also supports dictation across all apps. The […] industry https://arxiv.org/abs/2608.18661v1 sw-dd82c6ba0238ea54bda1 Wed, 19 Aug 2026 08:05:45 +0000 Streaming text-to-speech is essential for low-latency spoken dialogue systems, yet many systems wait for sentence-level text and are therefore only pseudo-streaming. True token-level synthesis must generate speech from uncertain prefixes while maintaining perceptual continuity over an unbounded stream with bounded cont research https://arxiv.org/abs/2608.18655v1 sw-a92d3eac07daf9a84946 Wed, 19 Aug 2026 08:00:40 +0000 The rapid progress in Artificial Intelligence has largely bypassed African languages, creating a digital divide that limits AI adoption on the continent. Recent open-source LLMs systematically underperform on African machine translation, while the lack of large-scale, high-quality, open-source parallel data has constra research https://techcrunch.com/2026/08/19/amazon-makes-its-ai-powered-alexa-free-on-fire-tv-no-prime-required sw-dcf28ade395637f24a69 Wed, 19 Aug 2026 15:00:00 +0000 Amazon is making its AI-powered Alexa+ assistant free on all compatible Fire TV devices in the U.S., automatically upgrading users whether or not they subscribe to Prime. industry https://inco.ai/blog/dflash2 sw-735a4492dd431c8ae49f Wed, 19 Aug 2026 20:28:43 +0000 developer https://arxiv.org/abs/2608.18628v1 sw-62c8d7757aceb5fba28f Wed, 19 Aug 2026 07:22:15 +0000 Aligned vision-language models (VLMs) are designed to balance grounded visual reasoning with safe generation behavior. However, we observe a striking phenomenon: under safety-constrained instruction, models frequently abstain from answering questions that remain correctly answerable under default instruction despite re research https://techcrunch.com/2026/08/19/calendly-throws-its-hat-into-meeting-note-taker-circus sw-bf0e5a0c7ddaf000891b Wed, 19 Aug 2026 14:09:39 +0000 Calendly is also releasing a meeting scheduling assistant called Callie. industry