https://megadose.ai/i/033de92f-6c71-4ab9-8096-89c9382740f7 Megadose en 2026-06-11T06:20:03+00:00 Sources: Alibaba replaces DingTalk CEO Chen Hang with Chen Yusen after an internal debate over the enterprise communication app's role in Alibaba's AI strategy (Bloomberg) https://megadose.ai/i/6bc04959-e664-4c85-8ba8-510453bbbbdc Megadose en 2026-06-11T01:50:01+00:00 Sources: OpenAI is considering drastically lowering its price for tokens in anticipation of similar cuts the company expects at Anthropic (Wall Street Journal) https://megadose.ai/i/cf9e0d67-91a2-431a-8c06-81a159a4080b Megadose en 2026-06-10T20:30:02+00:00 Oracle reports Q4 revenue up 21% YoY to $19.18B, vs. $19.1B est., and expects to raise ~$40B through a combination of debt and equity financing in FY 2027 (Juby Babu/Reuters) https://megadose.ai/i/86110c6a-fb84-4d52-80d4-bda8afda7528 Megadose en 2026-06-10T20:19:31+00:00 Fresh off bond sale, Amazon borrows $17.5 billion from banks as AI spending continues https://megadose.ai/i/575a9672-073d-4f96-b831-9bc80439fd07 Megadose en 2026-06-10T20:00:54+00:00 DiffusionGemma https://megadose.ai/i/c2c62f1d-191e-4d98-a88c-47cbf29b5674 Megadose en 2026-06-10T18:30:21+00:00 OpenAI and Visa partner to let AI agents make purchases online after users give their permission and to explore enterprise applications for AI-driven payments (Paige Smith/Bloomberg) https://megadose.ai/i/24eeec6b-b176-41f7-9325-60ca45b63a0f Megadose en 2026-06-10T17:59:57+00:00 Reroute, Don't Remove: Recoverable Visual Token Routing for Vision-Language Models https://megadose.ai/i/abfa7864-9a72-485b-920d-715fed61a741 Megadose en 2026-06-10T17:59:54+00:00 Context-Driven Incremental Compression for Multi-Turn Dialogue Generation https://megadose.ai/i/02e8455f-7d71-4d7d-86f3-c2da55f3d416 Megadose en 2026-06-10T17:59:35+00:00 FACTR 2: Learning External Force Sensing for Commodity Robot Arms Improves Policy Learning https://megadose.ai/i/d6005811-44c8-4c5e-9df5-9c745d2c7318 Megadose en 2026-06-10T17:58:49+00:00 DIRECT: When and Where Should You Allocate Test-Time Compute in Embodied Planners? https://megadose.ai/i/820d3abc-1db1-4a19-a77e-fe23d795ab08 Megadose en 2026-06-10T17:58:20+00:00 Doc-to-Atom: Learning to Compile and Compose Memory Atoms https://megadose.ai/i/35f241e3-4153-49de-a777-c48b8bbdcafd Megadose en 2026-06-10T17:52:15+00:00 TAHOE: Text-to-SQL with Automated Hint Optimization from Experience https://megadose.ai/i/487e7541-2190-4f56-8476-708adcf78acd Megadose en 2026-06-10T17:52:03+00:00 ATLAS: Active Theory Learning for Automated Science https://megadose.ai/i/8a14bce6-1045-44ec-b8c0-2d7ee6270552 Megadose en 2026-06-10T17:47:59+00:00 Which Models Are Our Models Built On? Auditing Invisible Dependencies in Modern LLMs https://megadose.ai/i/b762ad4b-1ad3-43a9-b753-d51e12729466 Megadose en 2026-06-10T17:47:07+00:00 APPO: Agentic Procedural Policy Optimization https://megadose.ai/i/c073c39c-e87e-4ae7-b845-a517bc77d82e Megadose en 2026-06-10T17:38:24+00:00 UniIntervene: Agentic Intervention for Efficient Real-World Reinforcement Learning https://megadose.ai/i/46b08e21-b020-4d0a-96d2-3f13040693cd Megadose en 2026-06-10T17:36:45+00:00 Breaking Entropy Bounds: Accelerating RL Training via MTP with Rejection Sampling https://megadose.ai/i/c7fd907e-cb6e-4e56-b51c-ffe8794ab7aa Megadose en 2026-06-10T17:34:12+00:00 Ambient Diffusion Policy: Imitation Learning from Suboptimal Data in Robotics https://megadose.ai/i/482b4530-08da-4a06-a3ac-9a743368a49f Megadose en 2026-06-10T17:33:55+00:00 On Subquadratic Architectures: From Applications to Principles https://megadose.ai/i/e127a807-9d65-4033-a096-9d61eae399e0 Megadose en 2026-06-10T17:31:34+00:00 Latent World Recovery for Multimodal Learning with Missing Modalities https://megadose.ai/i/6d33c8a3-388d-46f2-8d40-ee61054ce422 Megadose en 2026-06-10T17:31:16+00:00 Anatomy of Post-Training: Using Interpretability to Characterize Data and Shape the Learning Signal https://megadose.ai/i/5dc733c7-55e6-464b-a242-ad8eba2f9ea6 Megadose en 2026-06-10T17:26:08+00:00 CHORUS: Decentralized Multi-Embodiment Collaboration with One VLA Policy https://megadose.ai/i/6ab3578b-fea7-48af-88c4-686d57a63501 Megadose en 2026-06-10T17:17:52+00:00 Atlas H&E-TME: Scalable AI-Based Tissue Profiling at Expert Pathologist-Level Accuracy https://megadose.ai/i/af5e104c-0311-4680-83c8-b8c90683cbf5 Megadose en 2026-06-10T17:16:23+00:00 Claw-SWE-Bench: A Benchmark for Evaluating OpenClaw-style Agent Harnesses on Coding Tasks https://megadose.ai/i/1752a7bf-fafb-4ffb-b80c-1f7b2238d081 Megadose en 2026-06-10T17:15:28+00:00 ALIGNBEAM : Inference-Time Alignment Transfer via Cross-Vocabulary Logit Mixing https://megadose.ai/i/0633670e-3779-470d-b972-6d464e0f641c Megadose en 2026-06-10T17:05:50+00:00 Fourier Features Let Agents Learn High Precision Policies with Imitation Learning https://megadose.ai/i/75b5c1d0-612d-4815-82a0-0412c3f565b6 Megadose en 2026-06-10T17:04:59+00:00 Measuring Semantic Progress in Multi-turn Dialogue via Information Gain https://megadose.ai/i/24e056b6-0fbe-4aac-a3be-e90c67aba1bf Megadose en 2026-06-10T17:02:56+00:00 PROJECTMEM: A Local-First, Event-Sourced Memory and Judgment Layer for AI Coding Agents https://megadose.ai/i/9d27934c-3b9e-47d0-8056-1a848584df17 Megadose en 2026-06-10T16:35:04+00:00 Natural-Language Temporal Grounding in Hour-Long Videos is a Search Problem: A Benchmark and Empirical Decomposition https://megadose.ai/i/678c3285-1b61-48fd-a650-708a7e7a9935 Megadose en 2026-06-10T16:34:49+00:00 Learning What to Say to Your VLA: Mostly Harmless Vision Language Action Model Steering https://megadose.ai/i/f7bbc5dc-9608-4b2d-8036-082538ec30c1 Megadose en 2026-06-10T16:27:26+00:00 Measuring Epistemic Resilience of LLMs Under Misleading Medical Context https://megadose.ai/i/6710c80a-4bad-4de2-a074-dfdd00112864 Megadose en 2026-06-10T16:24:11+00:00 DiffusionGemma: 4x faster text generation https://megadose.ai/i/ea3364b9-91f7-4cb2-837f-b17ce5989f0f Megadose en 2026-06-10T16:23:37+00:00 SpikeDecoder: Realizing the GPT Architecture with Spiking Neural Networks https://megadose.ai/i/bbcc892b-2bcb-406e-a92c-3977ffcf187b Megadose en 2026-06-10T16:20:24+00:00 PianoKontext: Expressive Performance Rendering from Deadpan Context https://megadose.ai/i/cc1678b8-cf2c-491f-8f8e-24bfc8785963 Megadose en 2026-06-10T16:15:01+00:00 Google introduces DiffusionGemma, an experimental 26B-parameter open model that uses text diffusion for faster text generation compared to autoregressive models (The Keyword) https://megadose.ai/i/2d8ad655-6752-499e-a5b7-dad97a29e11c Megadose en 2026-06-10T16:14:23+00:00 Beyond Fully Random Masking: Attention-Guided Denoising and Optimization for Diffusion Language Models https://megadose.ai/i/bd6865b1-c8a9-47e4-a246-75a9c2e69624 Megadose en 2026-06-10T16:11:13+00:00 The Impossibility of Eliciting Latent Knowledge https://megadose.ai/i/7a3f31af-0e93-4ae7-a0a6-54ed0d5f2cf7 Megadose en 2026-06-10T16:09:37+00:00 DiffusionGemma: 4x Faster Text Generation https://megadose.ai/i/6ec83e41-82ef-486e-aca0-55ec26310b66 Megadose en 2026-06-10T16:05:06+00:00 Sources: Sam Altman told staff that OpenAI is expected to go public "within the next year" and OpenAI plans a tender offer "very soon" at a $687.69 share price (Stephanie Palazzolo/The Information) https://megadose.ai/i/e8693ee9-e58b-49b8-aa82-8d1a2347451b Megadose en 2026-06-10T15:52:24+00:00 Reassessing High-Performing LLMs on Polish Medical Exams: True Competence or Bias-Driven Performance? https://megadose.ai/i/507047ed-6662-40bc-b3cc-1371a10fa609 Megadose en 2026-06-10T15:30:54+00:00 Exclusive: OpenAI Preps New AI Model, Expects To Go Public ‘Within the Next Year’ https://megadose.ai/i/86a63164-103a-4d32-99b8-2ff6fb40fabd Megadose en 2026-06-10T15:25:55+00:00 CameraMatics, which uses AI to help fleet operators improve safety, reduce operational risk, and lower carbon emissions, raised €49M (Joe Brennan/The Irish Times) https://megadose.ai/i/978bec18-8ebf-41e2-9c8b-cccce60decbd Megadose en 2026-06-10T15:01:06+00:00 Apache Burr: Build reliable AI agents and applications https://megadose.ai/i/1b6e7170-0563-4494-b8a9-d0b8b58b5971 Megadose en 2026-06-10T15:00:51+00:00 Sources: Germany's Neura Robotics, which builds AI-powered humanoid robots, raised $1.4B from Tether, Qualcomm, Amazon, Nvidia, and others at a ~$7B valuation (Financial Times) https://megadose.ai/i/596fa348-869e-4c99-89ff-3166e186f5e9 Megadose en 2026-06-10T14:55:54+00:00 WMG acquires AI attribution startup Sureel, as it seeks to track when its songs and recordings are used in the training of AI models or in AI-generated works (Kristin Robinson/Billboard) https://megadose.ai/i/3195d8a2-a253-4c95-8e1f-a6e57d90d409 Megadose en 2026-06-10T14:31:30+00:00 Warner Music acquires AI attribution startup Sureel AI https://megadose.ai/i/80728501-5d5a-4626-a68a-a11b12b17911 Megadose en 2026-06-10T14:25:52+00:00 New York-based Jedify, whose platform connects to enterprises' knowledge sources via APIs to build a "context graph" for AI agents, raised a $24M Series A (Ram Iyer/TechCrunch) https://megadose.ai/i/fb7dcd81-122b-4b57-87d3-d78f9d7368d7 Megadose en 2026-06-10T14:05:02+00:00 Pi, whose "security brain" AI agent helps companies assess vulnerabilities to prioritize and patch, raised $35M at a $100M valuation; sources: xAI is a customer (Thomas Brewster/Forbes) https://megadose.ai/i/9e437d02-e1e6-4bce-991f-4a367eb5bd78 Megadose en 2026-06-10T14:00:53+00:00 Anthropic’s New Model Targets Power Users But Cuts Off AI Rivals https://megadose.ai/i/7dd26539-c699-4783-8d71-28f97f41f368 Megadose en 2026-06-10T13:51:31+00:00 Claude Code Managed Agents and model selector for Voice Mode https://megadose.ai/i/e62f0b6c-8650-43ed-8b44-86b67fb21022 Megadose en 2026-06-10T13:33:09+00:00 Jedify raises $24M to help companies arm AI agents with context on their business https://megadose.ai/i/9590867b-777a-4b42-ae82-6e93c50864a0 Megadose en 2026-06-10T13:15:02+00:00 Maneva, whose AI agents connect to existing camera infrastructure in factories to flag safety risks and measure worker productivity, raised a $27M Series A (Colin Campbell/Axios) https://megadose.ai/i/58f732af-60ab-4ed5-9fbd-d025b16b3e0d Megadose en 2026-06-10T13:07:56+00:00 Decart’s new world model can simulate hours of photorealistic driving — with some caveats https://megadose.ai/i/2f719640-8524-45b4-b137-5e875ef74ad5 Megadose en 2026-06-10T12:50:02+00:00 Cyera, an Israeli startup focused on protecting companies' data from AI-based threats, raised $600M at a $12B valuation, bringing its total funding to $2.3B (Niko Gallogly/New York Times) https://megadose.ai/i/f578a193-a7cb-413a-bca3-1d6a7b08bf5e Megadose en 2026-06-10T12:45:01+00:00 Poetic, which aims to use AI to automate tasks like financial compliance, emerges from stealth with $50M in funding from OpenAI and others at a $500M valuation (Paayal Zaveri/Bloomberg) https://megadose.ai/i/a18fec1e-3c32-4c2d-8711-ade49af3d376 Megadose en 2026-06-10T12:35:02+00:00 Hands-on with Siri AI: successfully executed multi-step prompts, understood context well, has strong guardrails, and is a bit more dispassionate than Gemini (Allison Johnson/The Verge) https://megadose.ai/i/5bbf0d5e-97fe-4482-a09d-9bd6a66054c4 Megadose en 2026-06-10T10:20:02+00:00 Las Vegas-based TensorWave, which offers AMD-powered cloud infrastructure, raised a $350M Series B led by AMD and Magnetar at a $1.55B post-money valuation (Robbie Whelan/Wall Street Journal) https://megadose.ai/i/7a4bd4c2-0451-4d5c-b30d-57b7846200d8 Megadose en 2026-06-10T09:55:30+00:00 Sources: SK Hynix plans to list its shares in the US as soon as August, seeking to capitalize on AI stock demand and broaden its investor base (Reuters) https://megadose.ai/i/7b8b99bf-3366-4c70-95df-88305b0e6574 Megadose en 2026-06-10T09:35:04+00:00 Swedish autonomous truck company Einride will debut on the Nasdaq today after completing a SPAC merger at a pre-money valuation of $1.35B (Dominic Chopping/Wall Street Journal) https://megadose.ai/i/a44c12c9-e70b-4e94-83da-ad7ceeee84da Megadose en 2026-06-10T07:05:26+00:00 Meta signs first AI data center deal in India with Reliance https://megadose.ai/i/5bd94cfa-e266-4d1d-bdee-4ae235964b99 Megadose en 2026-06-10T05:05:04+00:00 Sources: OpenAI is in advanced talks to lease a proposed 10GW data center campus in Ohio as part of a deal that could include financial backing from Nvidia (Anissa Gardizy/The Information) https://megadose.ai/i/d441048a-a749-4ee1-a4df-3f2c74dbb371 Megadose en 2026-06-10T04:55:15+00:00 Meta partners with Reliance to lease its first AI data center in India, with 168 MW capacity; separately, Meta says it has contracted ~1GW of renewable energy (Meta Newsroom) https://megadose.ai/i/d41bd44e-20e5-4ae6-a717-9f0c40752325 Megadose en 2026-06-10T03:50:21+00:00 [AINews] Anthropic Claude Fable 5 — Mythos but Safe, with Controversial Terms https://megadose.ai/i/b0d9e782-3d3e-4c9c-a46e-96fdd77e4f46 Megadose en 2026-06-10T02:30:01+00:00 Sources: SoftBank's talks with creditors to raise $6B through a margin loan backed by its OpenAI stake have stalled, weeks after it cut its target from $10B (Bloomberg) https://megadose.ai/i/742188bc-cb9f-4ca4-8db7-83505eb83211 Megadose en 2026-06-10T02:07:23+00:00 OpenAI in Talks to Lease 10 Gigawatt Ohio Data Center with Backing From Nvidia https://megadose.ai/i/985baa6d-4377-4fa5-827d-6001d85e3608 Megadose en 2026-06-10T01:55:09+00:00 Stepful, an AI-powered online training platform for healthcare workers, raised a $55M Series C led by Oak HC/FT, bringing its total funding to $105M (Heather Landi/Fierce Healthcare) https://megadose.ai/i/4fcd2ee4-c853-4e17-989f-081c6e7441a4 Megadose en 2026-06-10T00:00:00+00:00 Reason, Then Re-reason: Cross-view Revisiting Improves Spatial Reasoning https://megadose.ai/i/22cdbdd1-053c-41bc-b288-8b13861d593e Megadose en 2026-06-10T00:00:00+00:00 Redesign Mixture-of-Experts Routers with Manifold Power Iteration https://megadose.ai/i/b5f23a6c-0aa0-45dc-a81c-7e860bd3311d Megadose en 2026-06-10T00:00:00+00:00 Claw-SWE-Bench: A Benchmark for Evaluating OpenClaw-style Agent Harnesses on Coding Tasks https://megadose.ai/i/97b072bd-f454-4811-b45d-4396e8a587f7 Megadose en 2026-06-10T00:00:00+00:00 Toward Generalist Autonomous Research via Hypothesis-Tree Refinement https://megadose.ai/i/14a8ec47-919d-41cc-96fc-9983a189aba2 Megadose en 2026-06-10T00:00:00+00:00 Verifiable Environments Are LEGO Bricks: Recursive Composition for Reasoning Generalization https://megadose.ai/i/35627422-38aa-4e06-a83f-8006c9ec88b1 Megadose en 2026-06-10T00:00:00+00:00 InternVideo3: Agentify Foundation Models with Multimodal Contextual Reasoning https://megadose.ai/i/4cb4250e-2ba5-4afd-9918-101d00eb350c Megadose en 2026-06-10T00:00:00+00:00 World Model Self-Distillation: Training World Models to Solve General Tasks https://megadose.ai/i/8204e3b9-4dec-4c9a-acc0-692a23fb1732 Megadose en 2026-06-10T00:00:00+00:00 Breaking Entropy Bounds: Accelerating RL Training via MTP with Rejection Sampling https://megadose.ai/i/bcb0e228-c621-4c23-bd7f-a775b0edcfc5 Megadose en 2026-06-10T00:00:00+00:00 World Pilot: Steering Vision-Language-Action Models with World-Action Priors https://megadose.ai/i/81fd6ae7-9ebc-4086-9f03-5b15fa443b63 Megadose en 2026-06-10T00:00:00+00:00 ICA Lens: Interpreting Language Models Without Training Another Dictionary https://megadose.ai/i/dfd1b6bb-6e31-4cce-883e-480e37b9e3bc Megadose en 2026-06-10T00:00:00+00:00 On the Limits of LLM-as-Judge for Scientific Novelty Assessment https://megadose.ai/i/1600b1d7-213c-4165-9715-278ec482131f Megadose en 2026-06-09T23:59:54+00:00 Initial impressions of Claude Fable 5 https://megadose.ai/i/86b204d6-ac77-47a0-9ca3-b63858b89492 Megadose en 2026-06-09T23:25:01+00:00 CoreWeave founders have sold $2.3B in stock since the company's lockup period expired in August after a March 2025 IPO, reducing their combined holdings by ~25% (Ella Feldman/Bloomberg) https://megadose.ai/i/c76b70e6-5fc0-4a96-a4b6-f0510c8fc54b Megadose en 2026-06-09T22:55:02+00:00 Sources: SpaceX aims to launch orbital AI computing tests by late 2027 and has requested permission from regulators to launch up to 1M data-center satellites (Akash Sriram/Reuters) https://megadose.ai/i/01a30d22-ba25-4363-b6d6-3d2acd3f5a6b Megadose en 2026-06-09T22:05:01+00:00 Super Micro Computer aims to raise $7B in a series of equity and equity-linked financing transactions to fund its component purchases to satisfy AI orders (Harshita Mary Varghese/Reuters) https://megadose.ai/i/97b9874e-3b71-49b2-a0d4-295bb1dec5a4 Megadose en 2026-06-09T20:37:49+00:00 Anthropic’s Fable 5 can make weirdly fun video games with the click of a button https://megadose.ai/i/d5495b1f-f364-474e-9a62-30f478e08d2a Megadose en 2026-06-09T20:30:00+00:00 Anthropic says Fable 5 has invisible safeguards that use prompt modification, steering vectors, or PEFT to limit its effectiveness for frontier LLM development (Matthias Bastian/The Decoder) https://megadose.ai/i/83561dbe-9662-4c36-ac88-46973d170a9d Megadose en 2026-06-09T19:21:45+00:00 Ultrafast machine learning on FPGAs via Kolmogorov-Arnold Networks https://megadose.ai/i/825f903c-0f15-479c-9b61-887e35b672e2 Megadose en 2026-06-09T18:04:24+00:00 WWDC 2026: Everything announced on Siri AI, iOS 27, Apple Intelligence and more https://megadose.ai/i/0b534b92-f810-4fdd-9ca3-1f89a92300d0 Megadose en 2026-06-09T17:59:58+00:00 When to Align, When to Predict: A Phase Diagram for Multimodal Learning https://megadose.ai/i/ced4731d-2ede-4655-9fdf-dfbc83fddc75 Megadose en 2026-06-09T17:59:54+00:00 A Unifying Lens on Supervised Fine-Tuning Through Target Distribution Design https://megadose.ai/i/69052190-fd50-44e7-97ee-2988ced10472 Megadose en 2026-06-09T17:57:16+00:00 EEVEE: Towards Test-time Prompt Learning in the Real World for Self-Improving Agents https://megadose.ai/i/36760fe8-e2ee-47f4-87ad-311e84fad720 Megadose en 2026-06-09T17:57:16+00:00 EEVEE: Towards Test-time Prompt Learning in the Real World for Self-Improving Agents https://megadose.ai/i/0f3fa5c9-9b03-488b-a69a-09cca08e1b51 Megadose en 2026-06-09T17:56:36+00:00 Lip Forcing: Few-Step Autoregressive Diffusion for Real-time Lip Synchronization https://megadose.ai/i/3cc35807-d89b-474e-b546-ed219d87a979 Megadose en 2026-06-09T17:51:55+00:00 Data Journalist Agent: Transforming Data into Verifiable Multimodal Stories https://megadose.ai/i/cb7bf0bb-86e6-4609-b0fb-e1b629f9d339 Megadose en 2026-06-09T17:50:09+00:00 The Role of Feedback Alignment in Self-Distillation https://megadose.ai/i/c0cbeacc-5b32-4e1d-8e4d-5df2a7283ca1 Megadose en 2026-06-09T17:49:24+00:00 Predicting Future Behaviors in Reasoning Models Enables Better Steering https://megadose.ai/i/d5da51d7-1087-4ecc-8d11-ba78bdf65e2c Megadose en 2026-06-09T17:48:41+00:00 Piper: A Programmable Distributed Training System https://megadose.ai/i/f25ff405-2b30-4a94-8998-6fa4913c53b4 Megadose en 2026-06-09T17:46:55+00:00 Multi-Faceted Interactivity Alignment in Full-Duplex Speech Models https://megadose.ai/i/3d4e70d6-f62f-431b-a5e4-07bdc74c9d3b Megadose en 2026-06-09T17:44:23+00:00 ReasonAlloc: Hierarchical Decoding-Time KV Cache Budget Allocation for Reasoning Models https://megadose.ai/i/c524bff7-e843-4a55-b53d-4cf97803aaf0 Megadose en 2026-06-09T17:43:30+00:00 COGENT: Continuous Graph Emulators with Neural Ordinary Differential Equations for Long-Term Physical Forecasting https://megadose.ai/i/91e4e22f-e751-438b-b0b5-c501f3a6733d Megadose en 2026-06-09T17:39:22+00:00 Itô maps for any-step SDEs https://megadose.ai/i/10b7398d-b2a6-42d8-89f4-1cd9d4b0f6ac Megadose en 2026-06-09T17:35:02+00:00 Anthropic says Fable 5 uses conservative safety classifiers that trigger a fallback to Claude Opus 4.8 in ~5% of sessions, in areas like cybersecurity (Anthropic) https://megadose.ai/i/f343d9a4-9afd-4f27-99f1-090a5db0a96b Megadose en 2026-06-09T17:33:24+00:00 OncoTraj: a public benchmark for longitudinal resistance prediction in EGFR-mutant non-small-cell lung cancer on osimertinib https://megadose.ai/i/3cc89696-32b7-481d-9539-16dc4e018d99 Megadose en 2026-06-09T17:29:05+00:00 First-Order Trajectory Matching: Fast Ensemble Predictions of Chaotic, Turbulent, Stochastic Systems https://megadose.ai/i/a140dc29-1b52-45da-95ae-470c0a552eaa Megadose en 2026-06-09T17:24:27+00:00 Provenance-Grounded Gating and Adaptive Recovery in Synthetic Post-Training Data Curation https://megadose.ai/i/a2c6e3c3-a8b4-48cb-b764-4f2048d058c0 Megadose en 2026-06-09T17:16:03+00:00 TRACE: A Unified Rollout Budget Allocation Framework for Efficient Agentic Reinforcement Learning https://megadose.ai/i/627bc853-8949-4d29-814d-2f98d6d14625 Megadose en 2026-06-09T17:15:02+00:00 Anthropic prices both Claude Fable 5 and Mythos 5 at $10 per 1M input and $50 per 1M output tokens, less than half the price of Claude Mythos Preview (David Gewirtz/ZDNET) https://megadose.ai/i/d17e250c-3464-48f2-a641-dc87fdcfd895 Megadose en 2026-06-09T17:14:44+00:00 Towards Autonomous Accelerator Design: FPGA Accelerator Generation with SECDA https://megadose.ai/i/f8724bb4-75ad-4a2a-986f-5cebdd90ab89 Megadose en 2026-06-09T17:12:45+00:00 Anthropic releases Claude Fable 5, a "safe" Mythos-class model it says can't be used for cyberattacks, to the public, and Claude Mythos 5 to trusted orgs (Wired) https://megadose.ai/i/92bb1a61-f6d6-46be-ad6b-39625632c40a Megadose en 2026-06-09T17:03:37+00:00 FADA: Accessible fetal ultrasound interpretation and annotation with a selectively distilled unified vision-language model https://megadose.ai/i/88979a73-e650-4fcd-980d-9a7db9089d1e Megadose en 2026-06-09T17:03:19+00:00 PhantomBench: Benchmarking the Non-existential Threat of Language Models https://megadose.ai/i/3551fc93-765b-40b7-ac4d-c5a98d88ebc1 Megadose en 2026-06-09T17:02:27+00:00 Limitations of Learning Tanh Neural Networks with Finite Precision https://megadose.ai/i/4a715752-59be-4905-ae37-4a92fa43fa1f Megadose en 2026-06-09T17:00:00+00:00 Anthropic’s Claude Fable 5 is a version of Mythos the public can access today https://megadose.ai/i/7ccf8cf0-9a86-4891-80e7-7678ce639abf Megadose en 2026-06-09T16:57:10+00:00 Do Transformers Actually Help Intrusion Detection? A Temporal Sequence Evaluation on CIC-IDS2017 https://megadose.ai/i/1a96b28a-0ac4-4487-8418-3f8c439c15fa Megadose en 2026-06-09T16:51:13+00:00 RoboNaldo: Accurate, Stable and Powerful Humanoid Soccer Shooting via Motion-Guided Curriculum Reinforcement Learning https://megadose.ai/i/bcdde304-c478-4647-b621-1ab7ca3fabd6 Megadose en 2026-06-09T16:45:57+00:00 Test-Time Gradient Guidance of Flow Policies in Reinforcement Learning https://megadose.ai/i/5241f6a3-14f2-4233-aa1e-c818a6b005a0 Megadose en 2026-06-09T16:40:54+00:00 Unifying Local Communications and Local Updates for LLM Pretraining https://megadose.ai/i/87a7998e-467a-4ac8-a211-4cb4208d9d13 Megadose en 2026-06-09T16:39:32+00:00 VISTA: A Versatile Interactive User Simulation Toolkit for Agent Evaluation https://megadose.ai/i/1ad1249b-6a2a-465e-80ae-96cec5cb0057 Megadose en 2026-06-09T16:39:10+00:00 A History-Aware Visually Grounded Critic for Computer Use Agents https://megadose.ai/i/33a9b424-6566-492b-b743-9d5e49f20816 Megadose en 2026-06-09T16:36:54+00:00 Exploring the Design Space of Reward Backpropagation for Flow Matching https://megadose.ai/i/bd64a9a4-faa9-4c03-8057-a121d82b6115 Megadose en 2026-06-09T16:32:14+00:00 T1-Bench: Benchmarking Multi-Scenario Agents in Real-World Domains https://megadose.ai/i/4028fc48-cb74-423f-8718-cad3b240d9c4 Megadose en 2026-06-09T16:29:34+00:00 GRAFT: Gain-Recalibrated Adapters for Transformer-Based Neural Population Activity Modeling https://megadose.ai/i/f7b51133-14c0-4557-97d2-6ca451148c99 Megadose en 2026-06-09T16:29:30+00:00 Mora launches AI analytics platform with SQL transparency for teams https://megadose.ai/i/13c8a1c5-3012-4d38-b421-911156545c12 Megadose en 2026-06-09T16:20:36+00:00 Flexible Kernels for Protein Property Prediction https://megadose.ai/i/a9dd7f05-5faf-4aeb-af4e-f4b45a2e8ead Megadose en 2026-06-09T16:17:19+00:00 Attention Amnesia in Hybrid LLMs: When CoT Fine-Tuning Breaks Long-Range Recall, and How to Fix It https://megadose.ai/i/29d51513-74b1-4cb1-a993-97e7312e1346 Megadose en 2026-06-09T16:12:55+00:00 What Fits (Into Few Tokens) Doesn't Overfit: Compression and Generalization in ML Research Agents https://megadose.ai/i/55c86b83-32fe-4376-b04c-532299e46a39 Megadose en 2026-06-09T16:12:20+00:00 Generalized Conformal Predictive Systems Under Distributional Shifts https://megadose.ai/i/02b6ff09-4abc-404e-9e87-7b61b3aaaed8 Megadose en 2026-06-09T16:10:16+00:00 Workflow-GYM: Towards Long-Horizon Evaluation of Computer-use Agentic tasks in Real-World Professional Fields https://megadose.ai/i/9f96f0b5-f620-4522-9f9a-17fdeb9e47f9 Megadose en 2026-06-09T16:05:23+00:00 AuRA: Internalizing Audio Understanding into LLMs as LoRA https://megadose.ai/i/f4046f94-1249-4255-afc1-4b0431785456 Megadose en 2026-06-09T16:00:50+00:00 Broadcom to Help Finance Anthropic, OpenAI Chip Deals With Apollo, Blackstone https://megadose.ai/i/67a7ee20-7ca7-43f7-8b2b-ce23cd24ca0b Megadose en 2026-06-09T15:30:02+00:00 Google releases Gemini 3.5 Live Translate, which it says can deliver "near real-time speech-to-speech translation in over 70 languages" (The Keyword) https://megadose.ai/i/f22138cc-f9be-4888-af3b-693d7f20f831 Megadose en 2026-06-09T15:16:25+00:00 Fluid, natural voice translation with Gemini 3.5 Live Translate https://megadose.ai/i/a08439e8-3752-433e-9051-c735f82cbc30 Megadose en 2026-06-09T15:01:15+00:00 Can LLMs Beat Classical Hyperparameter Optimization Algorithms? https://megadose.ai/i/cef51d8f-f8d4-49be-ad62-b2c0601d09b8 Megadose en 2026-06-09T14:28:07+00:00 Role-Agent: Bootstrapping LLM Agents via Dual-Role Evolution https://megadose.ai/i/51b4e99b-61e5-46c8-8f3c-4ea3840d3487 Megadose en 2026-06-09T14:10:19+00:00 Introducing Gemma 4 12B: a unified, encoder-free multimodal model https://megadose.ai/i/4240f148-d540-4c04-9e67-abb7ec1e16a3 Megadose en 2026-06-09T14:06:47+00:00 Anthropic’s Mythos Is Coming Today; Apple Pursues Modest Goals with Siri Revamp https://megadose.ai/i/9f39559e-53eb-4d23-b313-d819a9d71216 Megadose en 2026-06-09T13:49:18+00:00 Maket debuts Auto-Complete for generating residential floor plans https://megadose.ai/i/16ce3fc2-7512-4324-9f58-7389b925d37a Megadose en 2026-06-09T13:47:25+00:00 Sandstone raises $30M to bring AI to in-house legal teams https://megadose.ai/i/ece92125-0920-42c5-a48e-e3615be976aa Megadose en 2026-06-09T13:27:03+00:00 Is Grep All You Need? How Agent Harnesses Reshape Agentic Search https://megadose.ai/i/5dbdeae4-5043-4982-a357-317f196ac224 Megadose en 2026-06-09T13:00:00+00:00 Lovable says it has hit $500M in annualized revenue, with 1 million new projects a week https://megadose.ai/i/2bd3ee35-492d-4472-820a-12b77d1eb9ca Megadose en 2026-06-09T12:30:02+00:00 Apple unveils new Apple Foundation Models: two on-device models, including a 20B-parameter multimodal model called AFM 3 Core Advanced, and three cloud models (Apple Machine Learning Research) https://megadose.ai/i/af96ee18-9007-4408-abdd-ca323f0daf1f Megadose en 2026-06-09T12:00:00+00:00 Claude Fable 5 and Claude Mythos 5 https://megadose.ai/i/ae1fd296-4e0f-4fc2-8363-0e2ffd8c93dc Megadose en 2026-06-09T11:50:28+00:00 When the Chain of Thought Knows Better: Failure Modes in Multi-Turn Reasoning Models https://megadose.ai/i/5b0b2af6-a408-4fbb-ac3c-085fbf6c33b7 Megadose en 2026-06-09T11:35:07+00:00 New York-based Standard Bots, which wants to make AI-powered robotic arms in the US, raised $200M led by General Catalyst and Robostrategy at a $1B valuation (Bloomberg) https://megadose.ai/i/9fbb492f-8bb1-4a9e-b735-9781714172b3 Megadose en 2026-06-09T10:10:00+00:00 Beacon Software, which acquires niche software businesses and transforms them with AI, raised a $225M Series C, bringing its total funding to $550M (Sarah Klearman/Wall Street Journal) https://megadose.ai/i/516b6fca-85cc-4e89-b4cb-856d878ed86c Megadose en 2026-06-09T10:05:01+00:00 France-based Alta Ares, which is building AI-powered air defense systems to counter drones and missiles, raised a €50M Series A led by Air Street Capital (Daphné Leprince-Ringuet/Sifted) https://megadose.ai/i/6a9ad478-bcf2-41e8-86f0-88310b96587a Megadose en 2026-06-09T09:56:51+00:00 How Does Reasoning Flow? Tracing Attention-Induced Information Flow for Targeted RL in LLMs https://megadose.ai/i/5f313eab-5bf4-46a5-819f-ed1f7ca307e0 Megadose en 2026-06-09T09:55:01+00:00 Sources: China is drafting plans to spend $295B over the next five years on building AI data centers, sourcing 80%+ of tech from local suppliers like Huawei (Charlie Zhu/Bloomberg) https://megadose.ai/i/ea623a1a-3e5e-446d-a953-f0b4e3261bbf Megadose en 2026-06-09T09:50:02+00:00 A group of Chinese tech companies, including Alibaba and CXMT, launches a ~$577M PE fund to boost China's "hard tech" sectors amid tightening US export curbs (Ann Cao/South China Morning Post) https://megadose.ai/i/be17b325-5c7a-4924-be76-49ffffe17aa2 Megadose en 2026-06-09T08:36:08+00:00 One Token per Multimodal Evidence: Latent Memory for Resource-Constrained QA