Aswath Damodaran, a New York University professor known as the 'Dean of Valuation,' said the market likely hit 'peak AI' months ago and predicted more consolidation and correction in the AI trade in t…
Microsoft Corp. paid a record $20 million in bug bounties to 562 researchers between July 1, 2025, and June 30, 2026, up from $17 million to 344 researchers the previous year, driven by expanded eligi…
Amazon, Google, and Microsoft reported surging cloud revenues and capital expenditures for Q2 2026, with Amazon raising its 2026 capex forecast to approximately $220 billion and Google raising its ran…
Morgan Stanley identified 24 AI-adopter stocks rated 'Overweight' that have seen relative net margins expand by 50 basis points in three months, with net margins now exceeding the broader market by ne…
OpenAI has released iMessage threads from former Apple engineer Chang Liu in response to Apple's trade secret lawsuit, showing Apple employees contacting Liu for technical help and internal files afte…
The UK Department for Business and Trade is considering requiring employers in Great Britain to consult workers before deploying workplace monitoring technology (WMT), including AI-powered productivit…
Researchers propose AdaMTP, an adaptive training paradigm for multi-token prediction that dynamically aligns prediction horizons with sequence predictability, outperforming standard MTP across mathema…
Researchers introduced Gaokerena, a family of compact Persian medical language models, with Gaokerena-V improving a translated medical MMLU benchmark from 46.28% to 49.31% and Gaokerena-R achieving 52…
Researchers introduced the Onramp-Sequence Cross-Distillation (OSCD), a post-training algorithm that improves native multilingual chain-of-thought reasoning in low-resource Southeast Asian languages b…
A study from arXiv (2608.00507v1) training a ~7M-parameter Transformer encoder on child-directed and read speech found that the learning objective, not architecture, determines perceptual narrowing in…
A new arXiv preprint (2608.00585v1) shows that per-chunk verification fails on multi-hop retrieval-augmented generation, with entailment scoring reaching only 0.643, 0.523, and 0.560 AUC on HotpotQA, …
Researchers propose SeDeM, a selective decompression framework that decouples compact memory storage from decoder conditioning for long-context question answering. On four long-context QA benchmarks, …
A new arXiv preprint (2608.00355v1) finds that apparent progress toward harder tasks in large language models is mostly a ceiling effect, but a smaller hard-task effect persists after controlling for …
A new arXiv paper (2608.00582v1) introduces BPE-guided insertion, a method to adapt byte-level BPE tokenizers to underrepresented languages without changing the model vocabulary size. On Ukrainian ada…
Researchers at Observatorio Lázaro have built a self-populating database that has recorded more than two million English lexical borrowings (anglicisms) in the Spanish digital press, processing 1.88 m…
A new arXiv preprint (2608.01014v1) introduces Cloud-ScPO, a topology-guided preference-mining framework that derives preference supervision from the hidden-state geometry of large language models (LL…
Researchers introduced TreeProbe, the first cultural-bias benchmark for Tibetan medicine, containing 4,719 expert-adjudicated items covering 467 diseases and 10 subtasks, and found that current large …
Researchers introduced MedUPSQA, a dataset of 21,874 mid-stream clinical decision points from 5,535 real case reports, and MedUPS, an alignment framework that uses reinforcement learning to improve ne…
Researchers propose RSTG (Recovering Learning Signals via Adaptive Teacher Guidance), a method that selectively applies on-policy distillation to negative zero-variance prompts in reinforcement learni…
Researchers introduced OoO-Spec, an out-of-order semantic speculation method that speeds up LLM tool calling by predicting function choices and argument values in parallel. Across seven targets and th…