https://megadose.ai/i/f4ccd709-44e9-40e9-aae5-7789d9ab0458 Megadose en 2026-07-04T12:52:08+00:00 Mistral releases Leanstral 1.5 open model for proof engineering https://megadose.ai/i/bc23719b-9cb4-47f2-b04f-4e9986c3353f Megadose en 2026-07-04T12:41:51+00:00 ICYMI: xAI debuts Grok Voice Agent Builder for Enterprises https://megadose.ai/i/4cb1075b-b484-4927-ab36-5d7e6bf4fe65 Megadose en 2026-07-04T12:30:01+00:00 Official data shows Hong Kong accounted for 50%+ of China's $239B in chip imports in the first five months of 2026, a record share, up from ~33% a decade ago (Bloomberg) https://megadose.ai/i/5add4af4-d7c2-4c68-a8d9-387a342cba67 Megadose en 2026-07-04T05:35:01+00:00 Micron breaks ground on its ~$9.3B Hiroshima factory expansion, part of its global ramp-up to meet AI demand, and plans to start HBM shipments from summer 2028 (Mari Kiyohara/Bloomberg) https://megadose.ai/i/37e48b58-dbb5-454c-8156-55acc61c01c0 Megadose en 2026-07-04T05:15:01+00:00 Singapore-based dConstruct Robotics, which develops spatial tech to let autonomous robots operate in complex, GPS-denied environments, raised a $125M Series A (Duc Dao/TNGlobal) https://megadose.ai/i/3b329d7d-e5dd-42d9-9b23-a1b0e06d6086 Megadose en 2026-07-04T05:10:01+00:00 Prague-based EquiLibre, which offers AI for quant hedge funds and is founded by three ex-Google DeepMind researchers, raised a Series A at a $500M valuation (Anna Heim/TechCrunch) https://megadose.ai/i/a3c4ce94-2a99-4495-a155-b36143f98069 Megadose en 2026-07-03T13:06:44+00:00 Condense launches proxy to cut AI coding agent bills by up to 72% https://megadose.ai/i/1c61aea7-aabd-434f-b805-707d4d491b32 Megadose en 2026-07-03T01:10:01+00:00 Sources: Anthropic's bankers have hired UK law firm Freshfields to advise on its IPO; it also advised on Google's acquisition of Wiz and ServiceNow's Armis deal (The Information) https://megadose.ai/i/6f521145-120f-41ed-9463-df9a1c25fdc3 Megadose en 2026-07-02T22:10:01+00:00 Sources: Crusoe is in active talks to raise ~$3B in a funding round expected to value the company in the ~$30B range, up from a ~$10B valuation in October (Bloomberg) https://megadose.ai/i/347147a7-2c7d-4f39-b392-79688b1fb722 Megadose en 2026-07-02T19:55:01+00:00 Vellum adds agent-to-agent AI collaboration for Slack https://megadose.ai/i/82f3a2be-22e0-4243-86c9-ce934b83f75c Megadose en 2026-07-02T19:15:01+00:00 Sources: ElevenLabs held early talks with investors on a secondary share sale for staff that would value the startup at ~$22B; it was valued at $11B in February (Bloomberg) https://megadose.ai/i/c7605157-a44b-4e00-b517-9fba7c1b31e1 Megadose en 2026-07-02T18:45:00+00:00 UK-based StirlingX, which develops secure data intelligence systems for defense and critical infrastructure, raised a $20M Series A, following a $11M seed (Cate Lawrence/Tech.eu) https://megadose.ai/i/79936bc9-608e-4c37-86ad-d449b5a2cb10 Megadose en 2026-07-02T18:44:02+00:00 Meta quietly launches vibe-coded gaming app Pocket https://megadose.ai/i/4a1f8205-b79c-4934-9a1f-7485b86b5430 Megadose en 2026-07-02T17:59:56+00:00 Distributed Attacks in Persistent-State AI Control https://megadose.ai/i/c02a8a16-f057-489b-bff3-9a2134fb0ebd Megadose en 2026-07-02T17:59:52+00:00 LACUNA: A Testbed for Evaluating Localization Precision for LLM Unlearning https://megadose.ai/i/1ddc2487-2f63-4743-988d-c5b65a6d7bd7 Megadose en 2026-07-02T17:59:50+00:00 Program-as-Weights: A Programming Paradigm for Fuzzy Functions https://megadose.ai/i/4498b910-e0e1-4ada-9a7c-0d62159fb2d0 Megadose en 2026-07-02T17:59:43+00:00 Online Safety Monitoring for LLMs https://megadose.ai/i/60d31923-70fe-412a-96b9-c37ccdad15c7 Megadose en 2026-07-02T17:59:26+00:00 ReContext: Recursive Evidence Replay as LLM Harness for Long-Context Reasoning https://megadose.ai/i/7dfa762a-bf14-4963-af9e-6ecda2017c42 Megadose en 2026-07-02T17:59:23+00:00 What LLM Agents Say When No One Is Watching: Social Structure and Latent Objective Emergence in Multi-Agent Debates https://megadose.ai/i/69947a47-cde2-48a9-8456-7e0f15dc4bfb Megadose en 2026-07-02T17:58:52+00:00 Reasoning LLM Improves Speaker Recognition in Long-form TV Dramas https://megadose.ai/i/ffbf7ebb-ac74-476d-8ae0-977e51826fa0 Megadose en 2026-07-02T17:58:29+00:00 DemoPSD: Disagreement-Modulated Policy Self-Distillation https://megadose.ai/i/15f0652d-c0a1-4ec4-98f5-db19c3e5f15e Megadose en 2026-07-02T17:57:31+00:00 Beyond Adam: SOAP and Muon for Faster, Label-Efficient Training of Machine Learning Interatomic Potentials https://megadose.ai/i/97011e39-1cb2-42ee-aca0-a8174bb502a1 Megadose en 2026-07-02T17:55:39+00:00 Controllable Sim Agents with Behavior Latents https://megadose.ai/i/da23a0dc-d3ff-4fa4-9ba0-8d93913be67d Megadose en 2026-07-02T17:55:24+00:00 Towards Robustness against Typographic Attack with Training-free Concept Localization https://megadose.ai/i/3db9dafb-714a-47be-a3b0-ce48ca869d57 Megadose en 2026-07-02T17:55:02+00:00 Memo: Microsoft is merging the consumer and enterprise versions of its Copilot chatbots into a single app featuring AI agents dubbed AutoPilot and coding tools (The Information) https://megadose.ai/i/25c21b2a-f3f6-4624-8135-2e3b077d71f3 Megadose en 2026-07-02T17:53:31+00:00 G-RRM: Guiding Symbolic Solvers with Recurrent Reasoning Models https://megadose.ai/i/59800e94-ff7a-40c2-af86-044b6295be46 Megadose en 2026-07-02T17:53:15+00:00 Visually Grounded Self-Reflection for Vision-Language Models via Reinforcement Learning https://megadose.ai/i/49551ad7-5752-4467-a2a3-79318f9d1574 Megadose en 2026-07-02T17:50:57+00:00 Combating Textual Noise and Redundancy: Entropy-Aware Dense Visual Token Pruning https://megadose.ai/i/987105a1-a1b0-4f76-ad03-2253ba5a41a8 Megadose en 2026-07-02T17:45:01+00:00 Tripo AI, which is developing 3D foundation models and world models for use in games, raised $150M, a month after it said it raised $200M (Dean Takahashi/GamesBeat) https://megadose.ai/i/ffade74f-908a-4341-8cb7-7829afda54b3 Megadose en 2026-07-02T17:35:20+00:00 TestEvo-Bench: An Executable and Live Benchmark for Test and Code Co-Evolution https://megadose.ai/i/cb266754-b286-45eb-aa3b-7bbeb112f7be Megadose en 2026-07-02T17:33:37+00:00 Learning to Move Before Learning to Do: Task-Agnostic pretraining for VLAs https://megadose.ai/i/f2e68434-357e-4466-9739-d0fe1f72d37b Megadose en 2026-07-02T17:30:38+00:00 Will Scaling Improve Social Simulation with LLMs? https://megadose.ai/i/1aabcc4c-4e95-45d9-9c7d-2667ac8f388e Megadose en 2026-07-02T17:27:34+00:00 OrbitQuant: Data-Agnostic Quantization for Image and Video Diffusion Transformers https://megadose.ai/i/34925337-d04b-4f52-b78a-b577dd3ef2bd Megadose en 2026-07-02T17:27:24+00:00 Neuron-Aware Data Selection for Annotation-Free LLM Self-Distillation https://megadose.ai/i/8d7449b3-8a26-40cd-9fc3-1047268d8230 Megadose en 2026-07-02T17:17:05+00:00 Understanding the Robustness of Distributed Self-Supervised Learning Frameworks Against Non-IID Data https://megadose.ai/i/105a9f4b-5c21-4120-bb5d-6774700ce01c Megadose en 2026-07-02T17:10:13+00:00 EvoPolicyGym: Evaluating Autonomous Policy Evolution in Interactive Environments https://megadose.ai/i/ca8f9942-ed6d-4521-a985-74e9f156d08c Megadose en 2026-07-02T17:09:14+00:00 Extreme Adaptive Transformer for Time Series Forecasting https://megadose.ai/i/348383e0-b420-4f99-bde0-a505cd843481 Megadose en 2026-07-02T17:08:21+00:00 Reasoning effort, not tool access, buys first-try reliability in agentic code generation: an observational study https://megadose.ai/i/59259130-a9cb-4950-b656-f4389341e03c Megadose en 2026-07-02T17:00:37+00:00 WorldSample: Closed-loop Real-robot RL with World Modelling https://megadose.ai/i/cf6b7189-9fa9-42ac-aa77-cb4cd4bfe3b5 Megadose en 2026-07-02T16:51:11+00:00 Neuron-Aware Active Few-Shot Learning for LLMs