https://megadose.ai/i/deafa2c1-0d12-49ab-a2ba-756f277e5242 Megadose en 2026-05-13T15:45:02+00:00 Sources: Arm and SoftBank expressed preliminary interest to acquire Cerebras weeks before its expected IPO; Cerebras rebuffed them (Bloomberg) https://megadose.ai/i/fea96715-1662-49ac-bdd4-2d53af0b34da Megadose en 2026-05-13T15:24:38+00:00 Cline releases open-source agent runtime SDK for coding agents https://megadose.ai/i/e0efd7fd-f7f9-44da-abc6-8a1660d3b2af Megadose en 2026-05-13T14:59:32+00:00 Amazon launches an AI shopping assistant for the search bar, powered by Alexa+ https://megadose.ai/i/d3e35540-b7a0-4740-929c-0ab9a9635a1d Megadose en 2026-05-13T14:30:01+00:00 Tencent says it plans to spend significantly more on AI infrastructure in H2 2026 as more China-designed AI chips become available "month by month" (The Information) https://megadose.ai/i/f9984b53-5f59-4959-b089-4e71d868be01 Megadose en 2026-05-13T14:29:19+00:00 Alibaba Expects $4.4 Billion in Annualized AI Revenue by Year-End https://megadose.ai/i/c01feae9-0cf7-40a9-96a0-40d89990507b Megadose en 2026-05-13T14:25:08+00:00 Investor docs: Anthropic's revenue run-rate is on track to hit $50B by the end of June; Ramp says more of its customers now use Anthropic than OpenAI, a first (Kate Clark/Wall Street Journal) https://megadose.ai/i/8f2deb16-ef0f-42e0-9611-74e7f583025e Megadose en 2026-05-13T14:15:09+00:00 UK chip startup Fractile raised a $220M Series B led by Factorial Funds, Accel and Founders Fund to make specialized logic and memory chips for inference (Robbie Whelan/Wall Street Journal) https://megadose.ai/i/98396ce3-fa7f-4ac0-9ca2-84b1fbb5cb08 Megadose en 2026-05-13T14:10:07+00:00 WhatsApp launches Incognito Chat, an AI chat mode built on Private Processing that Meta says lets users talk to AI without Meta being able to access the chats (Lily Hay Newman/Wired) https://megadose.ai/i/6204bc26-f8ed-43cd-9e23-048245da0fc9 Megadose en 2026-05-13T14:00:21+00:00 How Anthropic Could Take an OpenAI Supplier Off The Table https://megadose.ai/i/31753b3a-0136-4a14-9407-f76cf12f54a4 Megadose en 2026-05-13T13:57:42+00:00 thehype launches 24/7 AI-powered radio for founders https://megadose.ai/i/bf78715e-8ad2-4529-8662-ba6116ec20f6 Megadose en 2026-05-13T13:35:23+00:00 Nebius Reports 700% Increase in Q1 Revenue https://megadose.ai/i/cf96b3b1-1d49-4c0b-8553-7d5f768c5630 Megadose en 2026-05-13T13:27:09+00:00 Poppy debuts a proactive AI assistant to help organize your digital life https://megadose.ai/i/71c0d416-c3c1-41ac-a55b-66163de66364 Megadose en 2026-05-13T13:10:00+00:00 Amazon is replacing its Rufus AI shopping assistant with Alexa for Shopping, powered by Alexa+, on Amazon.com and its app for all customers in the US (Jennifer Pattison Tuohy/The Verge) https://megadose.ai/i/ee124f2b-be35-4d14-ad3f-78b3033e4a91 Megadose en 2026-05-13T13:00:32+00:00 TextGen is now a native desktop app. Open-source alternative to LM Studio (formerly text-generation-webui). https://megadose.ai/i/4f21ab97-bc14-4904-8490-d6a817c6da7e Megadose en 2026-05-13T12:11:33+00:00 Adaption aims big with AutoScientist, an AI tool that helps models train themselves https://megadose.ai/i/f1ce9f8f-350c-4e85-a68d-55e5a568b095 Megadose en 2026-05-13T09:45:01+00:00 SoftBank reports Q4 net income up 3x to ~$11.6B, above ~$1.5B est., driven by a $25B gain in its OpenAI stake; by October, its OpenAI investment will hit $64.6B (David Keohane/Financial Times) https://megadose.ai/i/0e5ddd01-a34c-4efe-a886-6b15859498ed Megadose en 2026-05-13T09:15:00+00:00 Richard Socher's Recursive Superintelligence raised $650M+ from GV, Greycroft, Nvidia, AMD, and others at a $4B valuation to pursue "recursive self-improvement" (Cade Metz/New York Times) https://megadose.ai/i/75592d73-544d-49b5-a934-1ac5df80252b Megadose en 2026-05-13T07:50:02+00:00 Anduril raised $5B led by Thrive Capital and a16z at a $61B valuation, up from $30.5B in June 2025, taking its total funding to $6.82B, and may IPO in 2027 (Sheera Frenkel/New York Times) https://megadose.ai/i/cd3a3039-4778-47d7-8691-324369cdfb58 Megadose en 2026-05-13T07:30:03+00:00 Former Alibaba Star Researcher Starts New AI Lab, Seeks $2 Billion Valuation https://megadose.ai/i/2a0603a1-26b5-46b8-a2dc-4870cc910e08 Megadose en 2026-05-13T05:00:14+00:00 Config, which is building a data layer for robotics foundation models, raised a $27M seed at a $200M+ valuation led by Samsung Venture Investment (Kate Park/TechCrunch) https://megadose.ai/i/b0ac26fb-bc76-41d4-b58e-8cc20de7d875 Megadose en 2026-05-13T04:45:01+00:00 Metis TechBio, a biotech company using AI to deliver and formulate drugs, surged as much as 185% in its Hong Kong market debut after raising ~$270M in its IPO (Amber Tong/Bloomberg) https://megadose.ai/i/43043b88-92c2-419b-a156-1ae749e44be2 Megadose en 2026-05-13T03:30:01+00:00 Paymentology, which provides real-time card and payment processing technology for banks and fintechs, raised $175M co-led by Apis Partners and Aspirity Partners (Tamara Djurickovic/Tech.eu) https://megadose.ai/i/8014e769-1a11-4281-887e-bdbbab1fc38b Megadose en 2026-05-13T02:55:01+00:00 Sources: Sam Altman recently discussed launching an AI compute company that is majority-owned by OpenAI, similar to Stargate's original data center initiative (Alex Heath/Sources) https://megadose.ai/i/febdbb60-d2b5-401c-82dd-3dfae90c8f22 Megadose en 2026-05-12T23:45:01+00:00 Sources: Anthropic is in talks to raise between $30B and $50B in a funding round that would value it at up to $950B (Mike Isaac/New York Times) https://megadose.ai/i/6a9c60b4-69e0-4e77-be6b-099264d02c53 Megadose en 2026-05-12T23:25:01+00:00 Vancouver-based quantum computing startup Photonic raised an additional $70M after a $130M raise announced in January, giving it a $2B post-money valuation (Josh Scott/BetaKit) https://megadose.ai/i/5b9e192e-aae9-467f-9947-6c3f5c55f9be Megadose en 2026-05-12T22:20:03+00:00 Source: Anthropic is in advanced talks to acquire New York-based Stainless, which helps developers generate SDKs from APIs, for at least $300M (The Information) https://megadose.ai/i/e2876a45-a1a8-44b2-a282-6d155b460898 Megadose en 2026-05-12T22:18:34+00:00 Google brings Gemini Intelligence automation to Android devices https://megadose.ai/i/3a5074de-110c-4ab5-929c-24944b4fbe5a Megadose en 2026-05-12T22:05:43+00:00 Meta to release Muse Spark in Voice Mode and Meta Glasses https://megadose.ai/i/07724259-cf4a-450c-acd8-59a1c2b7d459 Megadose en 2026-05-12T21:59:21+00:00 Anthropic in Talks to Buy Developer Tools Startup Used by OpenAI, Google https://megadose.ai/i/087e20e7-629d-4f49-a17b-abae04b8e5f1 Megadose en 2026-05-12T21:21:38+00:00 Sources: Anthropic is in early talks to raise at least $30B at a $900B+ valuation; the round is expected to close as soon as the end of this month (Bloomberg) https://megadose.ai/i/2dbecf2d-4574-4c05-975f-ff926691c4b7 Megadose en 2026-05-12T20:02:48+00:00 SAP Deepens Ties With Anthropic, Unveils AI Product Bundle https://megadose.ai/i/f8d8508d-fcda-4ab3-8ef3-08f018ae42ac Megadose en 2026-05-12T18:14:02+00:00 Google to Hire Hundreds of Engineers to Help Customers Adopt Its AI https://megadose.ai/i/47329f1b-00ab-4dbe-a170-6252d3d4bd92 Megadose en 2026-05-12T17:59:47+00:00 AlphaGRPO: Unlocking Self-Reflective Multimodal Generation in UMMs via Decompositional Verifiable Reward https://megadose.ai/i/8db2be24-e0ef-4d00-a8c2-e571e5026adf Megadose en 2026-05-12T17:59:34+00:00 LongMemEval-V2: Evaluating Long-Term Agent Memory Toward Experienced Colleagues https://megadose.ai/i/2cbbdbf6-8d34-4f1b-99ad-badf46061987 Megadose en 2026-05-12T17:59:34+00:00 Pion: A Spectrum-Preserving Optimizer via Orthogonal Equivalence Transformation https://megadose.ai/i/54bd27aa-9e9f-4bd5-aeb8-83ec60f57887 Megadose en 2026-05-12T17:59:26+00:00 Elastic Attention Cores for Scalable Vision Transformers https://megadose.ai/i/53b81dee-998a-452c-b094-5b55ba7e8d85 Megadose en 2026-05-12T17:58:27+00:00 Task-Adaptive Embedding Refinement via Test-time LLM Guidance https://megadose.ai/i/4f9ec876-641d-4882-98f1-a20c9be84e32 Megadose en 2026-05-12T17:58:20+00:00 Learning, Fast and Slow: Towards LLMs That Adapt Continually https://megadose.ai/i/43651030-4875-4c29-8d0a-913e2e74fc61 Megadose en 2026-05-12T17:57:48+00:00 Beyond GRPO and On-Policy Distillation: An Empirical Sparse-to-Dense Reward Principle for Language-Model Post-Training https://megadose.ai/i/32edcf32-1d1f-409c-ada6-948f91076206 Megadose en 2026-05-12T17:57:04+00:00 ToolCUA: Towards Optimal GUI-Tool Path Orchestration for Computer Use Agents https://megadose.ai/i/41dd1202-61a5-4b74-b97d-ae4fcc33bedc Megadose en 2026-05-12T17:56:59+00:00 OmniNFT: Modality-wise Omni Diffusion Reinforcement for Joint Audio-Video Generation https://megadose.ai/i/7463f128-0562-4c78-bb1b-e5299dfd012a Megadose en 2026-05-12T17:56:09+00:00 Needle: We Distilled Gemini Tool Calling Into a 26M Model https://megadose.ai/i/18187ae1-e9aa-4b17-a3cb-777e0ed7d657 Megadose en 2026-05-12T17:55:10+00:00 MEME: Multi-entity & Evolving Memory Evaluation https://megadose.ai/i/19543080-c238-4533-9511-80e291f80f7c Megadose en 2026-05-12T17:55:02+00:00 Routers Learn the Geometry of Their Experts: Geometric Coupling in Sparse Mixture-of-Experts https://megadose.ai/i/3c06572c-dd18-4739-a4c7-e0817a72ada2 Megadose en 2026-05-12T17:54:25+00:00 Reward Hacking in Rubric-Based Reinforcement Learning https://megadose.ai/i/c2300b05-28c9-45d9-afea-e35ec9719072 Megadose en 2026-05-12T17:53:47+00:00 KV-Fold: One-Step KV-Cache Recurrence for Long-Context Inference https://megadose.ai/i/c9a39501-85fc-4141-9fe3-052863cc2066 Megadose en 2026-05-12T17:51:26+00:00 Solve the Loop: Attractor Models for Language and Reasoning https://megadose.ai/i/bd554d28-6bb5-4969-8b93-7a4bc644b200 Megadose en 2026-05-12T17:50:08+00:00 Search Your Block Floating Point Scales! https://megadose.ai/i/5641e8be-2a67-4ae1-aef6-37eafee16d43 Megadose en 2026-05-12T17:48:45+00:00 Towards Affordable Energy: A Gymnasium Environment for Electric Utility Demand-Response Programs https://megadose.ai/i/b385a8fb-1420-4c28-8d41-3c56e9cd81e5 Megadose en 2026-05-12T17:48:09+00:00 A proximal gradient algorithm for composite log-concave sampling https://megadose.ai/i/67efce4c-fc2f-40b8-9d41-e10d2cf57979 Megadose en 2026-05-12T17:47:41+00:00 Multi-Stream LLMs: Unblocking Language Models with Parallel Streams of Thoughts, Inputs and Outputs https://megadose.ai/i/b5b18ef8-7eab-418f-b259-dd527baf7840 Megadose en 2026-05-12T17:44:41+00:00 TextSeal: A Localized LLM Watermark for Provenance & Distillation Protection https://megadose.ai/i/055ead65-b559-45a3-8c8b-8d12c9ece841 Megadose en 2026-05-12T17:43:49+00:00 Enabling AI-Native Mobility in 6G: A Real-World Dataset for Handover, Beam Management, and Timing Advance https://megadose.ai/i/d44e8c7b-1e08-4565-a4cf-46191bb77b92 Megadose en 2026-05-12T17:39:43+00:00 ORCE: Order-Aware Alignment of Verbalized Confidence in Large Language Models https://megadose.ai/i/6fa6e0d9-e9d9-439e-9386-c853b31bd890 Megadose en 2026-05-12T17:34:00+00:00 A Causal Language Modeling Detour Improves Encoder Continued Pretraining https://megadose.ai/i/bcf3b5a9-937f-40f8-aaff-c76c0e1df3f3 Megadose en 2026-05-12T17:32:15+00:00 CAAFC: Chronological Actionable Automated Fact-Checker for misinformation / non-factual hallucination detection and correction https://megadose.ai/i/e5ba1428-b524-4106-97d3-12ca68564ea5 Megadose en 2026-05-12T17:31:24+00:00 Google Weighs Using SpaceX to Launch Orbital Data Centers https://megadose.ai/i/7b246cce-f7e8-4d11-8632-c9c2b2954236 Megadose en 2026-05-12T17:31:00+00:00 Environment-Adaptive Preference Optimization for Wildfire Prediction https://megadose.ai/i/7a33f4b8-e516-4536-8a22-bf127cfa35fd Megadose en 2026-05-12T17:25:07+00:00 Anthropic announces 12 Claude plugins for the legal sector, including a "commercial counsel" tool for reviewing vendor agreements and a bar exam study tool (Rachel Metz/Bloomberg) https://megadose.ai/i/a9d2c917-6017-46d9-88f2-d79bcb4beea4 Megadose en 2026-05-12T17:23:30+00:00 Learning Minimally Rigid Graphs with High Realization Counts https://megadose.ai/i/07532fe0-22a0-4749-874c-a11e30eab6f3 Megadose en 2026-05-12T17:22:22+00:00 Geometric Factual Recall in Transformers https://megadose.ai/i/79467408-b1eb-490f-86b4-597bb88c7058 Megadose en 2026-05-12T17:16:30+00:00 Predicting Disagreement with Human Raters in LLM-as-a-Judge Difficulty Assessment without Using Generation-Time Probability Signals https://megadose.ai/i/c4d21329-858b-4424-9157-d0cae89098e5 Megadose en 2026-05-12T17:15:45+00:00 Formalize, Don't Optimize: The Heuristic Trap in LLM-Generated Combinatorial Solvers https://megadose.ai/i/2eeafb40-74f1-428d-b090-2c10fdfbc1ee Megadose en 2026-05-12T17:14:04+00:00 ORBIT: Preserving Foundational Language Capabilities in GenRetrieval via Origin-Regulated Merging https://megadose.ai/i/10dcd3a6-f342-4548-bd86-7c9e92b32dce Megadose en 2026-05-12T17:12:29+00:00 Aligning Flow Map Policies with Optimal Q-Guidance https://megadose.ai/i/de5ef1fe-dae1-4244-88e8-7a264899df2b Megadose en 2026-05-12T17:10:22+00:00 Show HN: Agentic interface for mainframes and COBOL https://megadose.ai/i/7fad102c-aa82-4245-aa9b-e71a770a3573 Megadose en 2026-05-12T17:09:41+00:00 Stories in Space: In-Context Learning Trajectories in Conceptual Belief Space https://megadose.ai/i/1babeaa1-99c4-46d5-bf35-afe15b10785f Megadose en 2026-05-12T17:09:32+00:00 Predicting Decisions of AI Agents from Limited Interaction through Text-Tabular Modeling https://megadose.ai/i/9fa84113-f0c4-435e-a784-cfd9935c5682 Megadose en 2026-05-12T17:05:59+00:00 Model-based Bootstrap of Controlled Markov Chains https://megadose.ai/i/b0c48750-b916-4f49-87cf-4a92da97f3f5 Megadose en 2026-05-12T17:03:26+00:00 Semantic Reward Collapse and the Preservation of Epistemic Integrity in Adaptive AI Systems https://megadose.ai/i/4c16ba15-19a5-4259-aab4-d31ce22686a9 Megadose en 2026-05-12T17:00:53+00:00 OGLS-SD: On-Policy Self-Distillation with Outcome-Guided Logit Steering for LLM Reasoning https://megadose.ai/i/39591427-a680-4dee-a8bb-2c891137aa0b Megadose en 2026-05-12T16:57:48+00:00 Detecting overfitting in Neural Networks during long-horizon grokking using Random Matrix Theory https://megadose.ai/i/86b9a6f6-fcff-4aa4-83d6-5f36f4b42e93 Megadose en 2026-05-12T16:56:19+00:00 Trajectory-Agnostic Asteroid Detection in TESS with Deep Learning https://megadose.ai/i/4016c723-349b-45ac-84f5-d0d4edd2aa2b Megadose en 2026-05-12T16:52:02+00:00 SEMIR: Semantic Minor-Induced Representation Learning on Graphs for Visual Segmentation https://megadose.ai/i/a32254bd-0c38-4fed-ae02-46de99522dbb Megadose en 2026-05-12T16:51:23+00:00 Events as Triggers for Behavioral Diversity in Multi-Agent Reinforcement Learning https://megadose.ai/i/6e4e7b94-272e-49e4-9af4-6fe7fab48d16 Megadose en 2026-05-12T16:47:40+00:00 Scalable Token-Level Hallucination Detection in Large Language Models https://megadose.ai/i/d49f9f0b-6c9c-4b1e-8522-d622b7d5e18f Megadose en 2026-05-12T16:45:38+00:00 Pretraining Exposure Explains Popularity Judgments in Large Language Models https://megadose.ai/i/f5720aab-36d4-4354-92ce-cf12836ef694 Megadose en 2026-05-12T16:44:47+00:00 Trust the Batch, On- or Off-Policy: Adaptive Policy Optimization for RL Post-Training https://megadose.ai/i/882b7b87-b2dc-46be-88f1-eb7a2828b6a9 Megadose en 2026-05-12T16:44:02+00:00 Discrete Flow Matching for Offline-to-Online Reinforcement Learning https://megadose.ai/i/412c23fa-dce0-45f5-8d0a-41416e0d9660 Megadose en 2026-05-12T16:34:36+00:00 Threads tests a Meta AI integration that works similarly to Grok https://megadose.ai/i/5d2c92fd-c01b-46f1-826d-4dd6189b4eb6 Megadose en 2026-05-12T16:10:01+00:00 Sources: Google is in talks with SpaceX and other companies for a rocket launch deal, as Google expands its own efforts to put orbital data centers in space (Wall Street Journal) https://megadose.ai/i/5071e549-08f4-44f6-a80c-de7b02d1f5e4 Megadose en 2026-05-12T15:40:00+00:00 Isomorphic Labs, an AI-powered drug discovery spinoff from Google DeepMind, raised $2.1B led by Thrive, after raising $600M in its first round in March 2025 (Pritam Biswas/Reuters) https://megadose.ai/i/05191c3c-ac9a-4d02-b081-c558ba2cc2fc Megadose en 2026-05-12T15:15:01+00:00 Exaforce, which uses AI agents to detect and thwart cyberattacks, raised a $125M Series B at a $725M valuation, bringing its total funding to $200M (Marina Temkin/TechCrunch) https://megadose.ai/i/54dc556b-e801-4857-8aea-f9aa2485c7a4 Megadose en 2026-05-12T15:10:01+00:00 Sources: Wispr AI, the developer of the voice dictation tool Wispr Flow, is in talks to raise ~$260M in a round that could more than double its valuation to $2B (Bloomberg) https://megadose.ai/i/589ccaf1-4ea5-47cb-b28f-c55268179c94 Megadose en 2026-05-12T14:50:01+00:00 SAP launches its new Autonomous Enterprise software suite, integrating data, cloud, AI, and automation features to streamline business processes (Mauro Orru/Wall Street Journal) https://megadose.ai/i/fcaf816f-1a87-4b4e-b09c-d91f98529d09 Megadose en 2026-05-12T14:33:12+00:00 TabPFN-3 just released: a pre-trained tabular foundation model for up to 1M rows [R][N] https://megadose.ai/i/63f90611-c771-47c9-9eff-4752a25c3d2d Megadose en 2026-05-12T13:01:06+00:00 SAP invests in German workflow automation platform n8n at a $5.2B valuation, up from $2.5B after an October 2025 equity raise, and agrees to embed n8n's tools (Yazhou Sun/Bloomberg) https://megadose.ai/i/89b3676e-e02b-4e8a-8f72-e406b96e03fa Megadose en 2026-05-12T12:00:50+00:00 AI voice startup Vapi raised a $50M Series B led by Peak XV, a source says at a $500M post-money valuation, after Amazon chose Vapi to handle 100% of Ring calls (Jagmeet Singh/TechCrunch) https://megadose.ai/i/475bf7b8-34e0-4867-931f-5ee55486247b Megadose en 2026-05-12T11:30:00+00:00 AI voice startup Vapi hits $500M valuation after winning Amazon Ring over 40 rivals https://megadose.ai/i/3fe668e0-095b-4ac9-ab44-9061a3da2d5e Megadose en 2026-05-12T10:20:01+00:00 Amp, which aims to buy excess computing capacity from data center operators to sell to startups, universities, and more, raised $1.3B from a16z and others (Cade Metz/New York Times) https://megadose.ai/i/1a6bf1e0-085b-40cc-a0d0-d2fb35233b61 Megadose en 2026-05-12T10:10:01+00:00 Sources: Jensen Huang was not invited to travel with President Trump on his China trip, a potential setback to Nvidia; Huang expressed his willingness to join (Bloomberg) https://megadose.ai/i/1135d7ff-287b-45bd-a91a-bee2424490ac Megadose en 2026-05-12T09:49:26+00:00 Unitree GD01: China's $537k rideable transformer robot is now in production https://megadose.ai/i/4ac96fa4-2be8-41f4-897e-46b9c87b8828 Megadose en 2026-05-12T01:40:00+00:00 Source: new revenue sharing terms cap OpenAI's payments to Microsoft at $38B; they previously could have totaled $135B through 2030 if long-term goals were met (The Information) https://megadose.ai/i/1ac260bd-2d42-4e03-8771-97b2472e5a0e Megadose en 2026-05-12T00:00:00+00:00 World Action Models: The Next Frontier in Embodied AI https://megadose.ai/i/0696343a-b56f-4e86-a4e9-1ed88640ea47 Megadose en 2026-05-12T00:00:00+00:00 L2P: Unlocking Latent Potential for Pixel Generation https://megadose.ai/i/a562d8cb-2469-40c8-af42-321b1a237ab8 Megadose en 2026-05-12T00:00:00+00:00 MEME: Multi-entity & Evolving Memory Evaluation https://megadose.ai/i/dfe1c306-fe40-4472-b9aa-308bd45bcb58 Megadose en 2026-05-12T00:00:00+00:00 Lite3R: A Model-Agnostic Framework for Efficient Feed-Forward 3D Reconstruction https://megadose.ai/i/eeecad26-45f5-4c34-8522-856e07a0dd90 Megadose en 2026-05-12T00:00:00+00:00 ToolCUA: Towards Optimal GUI-Tool Path Orchestration for Computer Use Agents https://megadose.ai/i/e5232326-f0bc-478b-93a9-2d6ba85a11e3 Megadose en 2026-05-12T00:00:00+00:00 AlphaGRPO: Unlocking Self-Reflective Multimodal Generation in UMMs via Decompositional Verifiable Reward https://megadose.ai/i/0df27d2d-d07f-4b8a-95bc-12d0cadf69e2 Megadose en 2026-05-12T00:00:00+00:00 Geometric Factual Recall in Transformers https://megadose.ai/i/6fb1f4a2-8bae-4509-922c-c953a0af388a Megadose en 2026-05-12T00:00:00+00:00 VidSplat: Gaussian Splatting Reconstruction with Geometry-Guided Video Diffusion Priors https://megadose.ai/i/35ebdcfe-fbf4-4faf-b4b7-da64ecb97d3b Megadose en 2026-05-12T00:00:00+00:00 LongMemEval-V2: Evaluating Long-Term Agent Memory Toward Experienced Colleagues https://megadose.ai/i/b99c82c8-f9af-4d5b-b4fd-8845e338ecff Megadose en 2026-05-12T00:00:00+00:00 Do Enterprise Systems Need Learned World Models? The Importance of Context to Infer Dynamics https://megadose.ai/i/4853e2e7-46e0-4ab8-a387-9be2925e3493 Megadose en 2026-05-12T00:00:00+00:00 SenseNova-U1: Unifying Multimodal Understanding and Generation with NEO-unify Architecture https://megadose.ai/i/5e2d0528-022f-4495-81cb-b8e1edaed051 Megadose en 2026-05-12T00:00:00+00:00 Reward Hacking in Rubric-Based Reinforcement Learning https://megadose.ai/i/7ff919db-d3f7-4cef-920a-5430747dc836 Megadose en 2026-05-12T00:00:00+00:00 From Web to Pixels: Bringing Agentic Search into Visual Perception https://megadose.ai/i/ca99d6d0-0bd5-48e0-97e6-24f3264b0d7b Megadose en 2026-05-12T00:00:00+00:00 Beyond the Last Layer: Multi-Layer Representation Fusion for Visual Tokenization https://megadose.ai/i/75d9fb26-d29f-440f-bc24-51641383b0ab Megadose en 2026-05-12T00:00:00+00:00 Images in Sentences: Scaling Interleaved Instructions for Unified Visual Generation https://megadose.ai/i/524b4b6b-4459-4b30-9a07-2c1244ce4cb4 Megadose en 2026-05-12T00:00:00+00:00 Beyond GRPO and On-Policy Distillation: An Empirical Sparse-to-Dense Reward Principle for Language-Model Post-Training https://megadose.ai/i/92eee1ac-8532-449d-88d1-afa2fa266351 Megadose en 2026-05-12T00:00:00+00:00 δ-mem: Efficient Online Memory for Large Language Models https://megadose.ai/i/37b42bb3-00bc-41e0-af48-e7fdc5ecbad9 Megadose en 2026-05-12T00:00:00+00:00 CausalCine: Real-Time Autoregressive Generation for Multi-Shot Video Narratives https://megadose.ai/i/2f38e15b-6702-42aa-a928-28fe2d1b939a Megadose en 2026-05-12T00:00:00+00:00 LychSim: A Controllable and Interactive Simulation Framework for Vision Research https://megadose.ai/i/41342eb7-9a76-4fa2-ba11-a386369ebb4d Megadose en 2026-05-11T23:00:00+00:00 OpenAI launches Daybreak, a cybersecurity initiative integrating AI models and Codex Security to help organizations patch vulnerabilities (Alexey Shabanov/TestingCatalog AI News) https://megadose.ai/i/520463bf-8c0e-4d32-92f1-fdd49a948ed8 Megadose en 2026-05-11T22:50:01+00:00 Thinking Machines Lab details interaction models, which can think and respond in real time, letting users and AI interact continuously for better collaboration (Thinking Machines Lab) https://megadose.ai/i/c45d7487-a4ea-4617-930e-9d8386fd730d Megadose en 2026-05-11T22:04:00+00:00 Thinking Machines announced new Interaction Voice Models https://megadose.ai/i/bf813e53-1b96-431a-94b2-2ec81efdf3cb Megadose en 2026-05-11T21:48:52+00:00 OpenAI announces Daybreak initiative around Codex Security https://megadose.ai/i/f4f57ab8-ee62-41eb-befd-62fe5ffa7838 Megadose en 2026-05-11T21:16:09+00:00 Anthropic adds Agent View to Claude Code CLI interface https://megadose.ai/i/92580a01-5ae0-41c3-a157-e9dce8b98122 Megadose en 2026-05-11T17:59:29+00:00 ELF: Embedded Language Flows https://megadose.ai/i/98413a98-d279-4c32-93a6-612139048761 Megadose en 2026-05-11T17:58:45+00:00 Variational Inference for Lévy Process-Driven SDEs via Neural Tilting https://megadose.ai/i/aafed373-d35e-47af-b057-476f5d4b5715 Megadose en 2026-05-11T17:55:13+00:00 Dynamic Skill Lifecycle Management for Agentic Reinforcement Learning https://megadose.ai/i/157aebc5-391f-476c-bc72-64ede366f14e Megadose en 2026-05-11T17:52:15+00:00 Optimal and Scalable MAPF via Multi-Marginal Optimal Transport and Schrödinger Bridges https://megadose.ai/i/0bd619f0-b7c2-4b4e-bb7d-39b9fff93407 Megadose en 2026-05-11T17:50:51+00:00 Shepherd: A Runtime Substrate Empowering Meta-Agents with a Formalized Execution Trace https://megadose.ai/i/f42f23e4-c752-43f5-a286-9ddccec9f71a Megadose en 2026-05-11T17:49:43+00:00 WildClawBench: A Benchmark for Real-World, Long-Horizon Agent Evaluation https://megadose.ai/i/95cb4c57-d085-48f3-bda7-326c170c2d2a Megadose en 2026-05-11T17:49:28+00:00 Equivariant Reinforcement Learning for Clifford Quantum Circuit Synthesis https://megadose.ai/i/e4432e5b-c658-4bc3-be48-74a5745e5ea6 Megadose en 2026-05-11T17:49:09+00:00 Revisiting Policy Gradients for Restricted Policy Classes: Escaping Myopic Local Optima with $k$-step Policy Gradients https://megadose.ai/i/31407ddb-e519-4df5-bc6a-f788f708172f Megadose en 2026-05-11T17:46:33+00:00 Engineering Robustness into Personal Agents with the AI Workflow Store https://megadose.ai/i/405bde03-632f-41c2-ae05-1171e26934f8 Megadose en 2026-05-11T17:46:24+00:00 DataMaster: Towards Autonomous Data Engineering for Machine Learning https://megadose.ai/i/c8b92ecd-3935-476a-899a-d353e0428d86 Megadose en 2026-05-11T17:41:38+00:00 Beyond Red-Teaming: Formal Guarantees of LLM Guardrail Classifiers https://megadose.ai/i/506c7408-411c-4828-95b0-87740d4d4daf Megadose en 2026-05-11T17:40:38+00:00 RubricEM: Meta-RL with Rubric-guided Policy Decomposition beyond Verifiable Rewards https://megadose.ai/i/2d49d11b-1934-4a18-8069-5e76da726ec1 Megadose en 2026-05-11T17:38:17+00:00 V4FinBench: Benchmarking Tabular Foundation Models, LLMs, and Standard Methods on Corporate Bankruptcy Prediction https://megadose.ai/i/35c99ddb-831a-4dd1-969f-0d96e39c4774 Megadose en 2026-05-11T17:35:10+00:00 Grounded or Guessing? LVLM Confidence Estimation via Blind-Image Contrastive Ranking https://megadose.ai/i/59d4f828-5f1a-40db-b22b-5b055068a29a Megadose en 2026-05-11T17:33:28+00:00 Unmasking On-Policy Distillation: Where It Helps, Where It Hurts, and Why https://megadose.ai/i/2cf2f403-6915-4f42-b71e-3cc2c9e96840 Megadose en 2026-05-11T17:33:01+00:00 Shields to Guarantee Probabilistic Safety in MDPs https://megadose.ai/i/8a3efb58-8121-4d05-ae90-9534c352fee4 Megadose en 2026-05-11T17:32:29+00:00 LoKA: Low-precision Kernel Applications for Recommendation Models At Scale https://megadose.ai/i/d7206993-d2f9-44a3-ac60-c58b9c1ee4f1 Megadose en 2026-05-11T17:27:16+00:00 AssayBench: An Assay-Level Virtual Cell Benchmark for LLMs and Agents https://megadose.ai/i/66fa1a4c-686d-475d-8c1a-82eb569e0b30 Megadose en 2026-05-11T17:27:15+00:00 Compute Where it Counts: Self Optimizing Language Models https://megadose.ai/i/6d9385a6-1a23-42ba-8c55-5d6588433382 Megadose en 2026-05-11T17:25:47+00:00 CADBench: A Multimodal Benchmark for AI-Assisted CAD Program Generation https://megadose.ai/i/947be4fc-582c-441f-82e3-fa494b18da1e Megadose en 2026-05-11T17:24:42+00:00 Attractor-Vascular Coupling Theory: Formal Grounding and Empirical Validation for AAMI-Standard Cuffless Blood Pressure Estimation from Smartphone Photoplethysmography https://megadose.ai/i/74915b80-b54a-413e-ac74-98994c52689f Megadose en 2026-05-11T17:20:58+00:00 Remember the Decision, Not the Description: A Rate-Distortion Framework for Agent Memory https://megadose.ai/i/523e0edb-43f6-4542-bbf2-d2e09feb0c46 Megadose en 2026-05-11T17:17:02+00:00 BEACON: A Multimodal Dataset for Learning Behavioral Fingerprints from Gameplay Data https://megadose.ai/i/877c1060-b901-46f3-b45b-24380117a8af Megadose en 2026-05-11T17:13:36+00:00 BenchCAD: A Comprehensive, Industry-Standard Benchmark for Programmatic CAD https://megadose.ai/i/8b523793-e4b5-45c9-9fd9-85deccc2bd3a Megadose en 2026-05-11T17:10:44+00:00 DGPO: Beyond Pairwise Preferences with Directional Consistent Groupwise Optimization https://megadose.ai/i/1beafe56-a78d-41d9-8aa5-9256131fb689 Megadose en 2026-05-11T17:10:35+00:00 RUBEN: Rule-Based Explanations for Retrieval-Augmented LLM Systems https://megadose.ai/i/0486af31-f6bd-4dc3-8bb6-0aab4bebb348 Megadose en 2026-05-11T17:05:52+00:00 Masked Generative Transformer Is What You Need for Image Editing https://megadose.ai/i/d43a317d-f311-4eb7-92be-be5a6311c1b4 Megadose en 2026-05-11T17:02:09+00:00 Learning More from Less: Exploiting Counterfactuals for Data-Efficient Chart Understanding https://megadose.ai/i/b32ef872-d840-42b9-8418-67bcb6b04f9d Megadose en 2026-05-11T17:00:18+00:00 The Generalized Turing Test: A Foundation for Comparing Intelligence https://megadose.ai/i/23db024f-8330-4c65-bbdd-7f0844a68e91 Megadose en 2026-05-11T16:58:57+00:00 Rethinking Agentic Search with Pi-Serini: Is Lexical Retrieval Sufficient? https://megadose.ai/i/b2a9a7ef-5986-4804-a921-d2e43b1adfc2 Megadose en 2026-05-11T16:58:48+00:00 Conditional anomaly detection methods for patient-management alert systems https://megadose.ai/i/a303e5b3-c2b7-4b20-9dde-31d168458523 Megadose en 2026-05-11T16:56:44+00:00 BabelDOC: Better Layout-Preserving PDF Translation via Intermediate Representation https://megadose.ai/i/df815097-c742-4219-94ef-7dfc8e83e6f6 Megadose en 2026-05-11T16:55:16+00:00 Training-Free Cultural Alignment of Large Language Models via Persona Disagreement https://megadose.ai/i/7d5dae1c-db8a-4dd6-b821-7c133b273246 Megadose en 2026-05-11T16:54:23+00:00 Clin-JEPA: A Multi-Phase Co-Training Framework for Joint-Embedding Predictive Pretraining on EHR Patient Trajectories https://megadose.ai/i/b3d69b67-c1fa-4d80-a9e2-8b25ce611802 Megadose en 2026-05-11T16:40:01+00:00 Sources: German defense tech startup Helsing is set to raise $1.2B led by Dragoneer and Lightspeed at a valuation of about $18B, up from $14B in June 2025 (Financial Times)