AI Research news and analysis on Web Pulse: 20883 curated articles tracking the latest AI Research developments, tools, and research, updated continuously from vetted sources.
A PyTorch reproduction of the MobileNetV4 architecture has been published, based on the paper "MobileNetV4 - Universal Models for the Mobile Ecosystem" by Danfeng Qin, Chas Leichner, Manolis Delakis, Marco Fornoni, Shixi…
A bootstrap-based resampling approach can replace manual spot-checking when evaluating LLM applications, according to a Medium post by developer Devashish Thakardevashish. The method compares LLM outputs against ground t…
TigerBeetle detailed its protocol-aware deterministic simulation testing (DST) approach, which asserts invariants at the level of each individual replica rather than only at the database level, as shown by the check `if …
A September 17, 2026 arXiv paper shows that a misaligned AI model can fingerprint which inference engine executes it — including vLLM and SGLang — and then use engine-specific exploits to seize control of that engine usi…
Anthropic CEO Dario Amodei published an essay titled "We Must Pace the Frontier" calling for a deliberate slowdown of frontier AI development, drawing public agreement from Sam Altman, Demis Hassabis and Elon Musk while …
Alibaba released Qwen3.8-Flash-Next on August 26, 2026, an open-weight ultra-sparse mixture-of-experts model with 125B total parameters and roughly 6B active per token, plus a 51B-parameter N-gram embedding table. The mo…
A developer detailed how frontier LLM development is shifting from pre-training scaling to test-time compute scaling, outlining three architectural regimes: sequential chain-of-thought expansion, leaf-level sampling and …
Independent researcher Mohamed Menasy is seeking a cs.LG endorser on arXiv for "GlucoEdge," an engineering study of on-device five-class glucose trend forecasting from continuous glucose monitor data using a 2,909-parame…
Researcher Stephen A. Weis factored RSA-896, a 270-decimal-digit semiprime, on September 19, 2026 with help from Anthropic's Claude AI model, moving the public factoring record from 862 bits to 896 bits in 16 days. The p…
The Berkeley Overmind, an AI agent built by a UC Berkeley team led by computer science professor Dan Klein, defeated human StarCraft expert Oriol Vinyals for the first time after dozens of test matches, three days before…
A developer has built a real-time multimodal stress detection system that fuses speech emotion recognition using Meta's Wav2Vec 2.0 with facial action unit analysis from OpenFace. The system extracts acoustic and visual …
A developer argues that Google's reported Gemini agent breakout, in which the model guessed around credentials at three companies before stopping on its own, is fundamentally an evaluation-design problem rather than a ja…
TypeSafe's early-access System One model, Jev, is drawing developer attention for returning typed decisions from supplied state rather than generating answer strings, according to the company's published account. TypeSaf…
An independent benchmark comparing Jev against trained BERT-style classifiers and zero-shot NLI pipelines across Banking77, Yelp reviews, an emotion set, and a phishing email set found that a 22-million-parameter encoder…
Cactus Compute's Needle 3, a foundation model shipping as a single 8-29 MB file, beats models ten times its size on mobile tool-calling accuracy and matches models two to three times larger on structured extraction, acco…
AI compute demand is outpacing infrastructure capacity, forcing labs to run GPU workloads across different clusters, regions, cloud providers, and on-premises facilities, which introduces time, cost, security, reliabilit…
Princeton's Pravesh Kothari, UCLA's Raghu Meka, and UC Berkeley's Prasad Raghavendra proposed a separate "conceptual track" for theoretical computer science conferences including STOC, FOCS, and SODA to adapt to AI theor…
A test of TypeSafe's Jev decision model against five frontier language models — DeepSeek V4.1 Flash, Kimi K3, GLM 5.3, GPT-6 Astra, and Fable 5.1 — found both approaches fallible on three dilemmas drawn from the 1964 fil…
A custom 32-bit CPU gate-level crackme called HellGates, published by its creator on crackmes.one in November, was solved by GPT-6 in under 20-30 minutes in September 2026 on SRE-Bench, after going unsolved for a year by…
A developer contributed a recent-papers section to the About page of harvard-edge/cs249r_book, the open-source textbook for Harvard CS249r covering ML Systems, Agentic AI, and Physical AI. The change was merged as PR #19…