Skip to content
gentic.news — AI News Intelligence Platform
Connecting to the Living Graph…
gentic.news

AI Research

Breaking AI research news: latest papers from arXiv, NeurIPS, ICML, and top labs. Track transformer architecture advances, reasoning breakthroughs, and scientific discoveries in machine learning and artificial intelligence.

AI ResearchFunding & BusinessProducts & LaunchesBig TechStartupsOpen SourcePolicy & EthicsOpinion & Analysis
A computer screen displays code from Andrej Karpathy's autoresearch repository, with a robotic arm icon symbolizing…
AI Research
75

AI Agents Now Training Other AI Models, Sparking Autoresearch Trend

AI agents are now being used to train other AI models, creating advanced agentic systems. This development stems from Andrej Karpathy's autoresearch repository and represents early-stage automation of AI research.

x.com/Apr 21, 2026/3 min read
open sourceautomationmachine learning
A collage of dozens of small AI-generated images in varied artistic styles, from oil painting to sketch, arranged in…
AI Research
85

Tencent Releases MegaStyle: 1.4M AI-Generated Image Style Dataset

Tencent has open-sourced MegaStyle, a 1.4 million image dataset for style transfer and text-to-image fine-tuning. It was generated by systematically pairing 170,000 style prompts with 400,000 content prompts using the Qwen-Image model.

x.com/Apr 21, 2026/3 min read
open sourcecomputer visiondatasets
Xiaomi's autonomous driving system OneVL processes road scenes via latent Chain-of-Thought reasoning, outperforming…
AI Research
95

Xiaomi's OneVL Uses Latent CoT to Beat Explicit CoT in Autonomous Driving

Xiaomi's Embodied Intelligence Team released OneVL, a vision-language model using latent Chain-of-Thought reasoning. It achieves state-of-the-art results on four autonomous driving benchmarks without the latency penalty of explicit reasoning steps.

x.com/Apr 21, 2026/3 min read
multimodal-airesearchautonomous-vehicles
Architecture diagram of the CGCMA model showing news text flowing into a transformer encoder and merging with crypto…
AI Research
85

CGCMA Model Achieves +0.449 Sharpe Ratio in Asynchronous Crypto News Fusion

Researchers propose CGCMA, a model for fusing sporadic news with continuous market data. It achieved a +0.449 Sharpe ratio on a new crypto trading benchmark, showing gains not explained by simple heuristics.

arxiv.org/Apr 21, 2026/3 min read/Multi-Source
finance-aimultimodal-airesearch
A person types on a laptop with code on screen, a glowing brain icon hovers above the keyboard, symbolizing AI reasoning
AI Research
100

Kimi 2.6 Thinking Shows Promise as Open Weights Model, Lags Behind Closed SoTA

An initial evaluation of Moonshot AI's Kimi 2.6 Thinking model finds it generates extensive reasoning traces but delivers only 'okay-ish' results on creative and coding tasks, highlighting the persistent open vs. closed model gap.

x.com/Apr 21, 2026/3 min read/Widely Reported
open sourcereasoningmodel release
NVIDIA researcher pointing at a diagram of chip design flow on a screen, with AI-generated optimization suggestions…
AI Research
85

NVIDIA Research Shows AI Can Optimize Decades-Old EDA Tools Like ABC

New NVIDIA research indicates AI can be used to optimize Electronic Design Automation (EDA) tools, such as the classic ABC system, which have been manually tuned by engineers for decades. This could automate a core, labor-intensive bottleneck in semiconductor design.

x.com/Apr 21, 2026/3 min read
researchmachine learningsemiconductors
A document database icon with a skull symbol overlaid, representing poisoned data corrupting a RAG AI system
AI Research
85

Poisoned RAG: 5 Documents Can Corrupt 'Hallucination-Free' AI Systems

Researchers proved that planting a handful of poisoned documents in a RAG system's database can cause it to generate confident, incorrect answers. This exposes a critical vulnerability in systems marketed as 'hallucination-free'.

x.com/Apr 20, 2026/3 min read
hallucinationssecurityresearch
A bar chart comparing AI agents and human teams on economic analysis, showing AI results clustered near the median…
AI Research
85

AI Agents Show Consistent Economic Analysis, Reducing Human Disagreement

A new study finds AI agents like Claude Code and Codex produce economic analyses with far less disagreement than human teams, landing near the human median but with no extreme outliers. This indicates AI's potential for scalable, consistent research support.

x.com/Apr 20, 2026/3 min read
economicslarge language modelsai research
Developer coding on a laptop with AI assistant interface displaying code, symbolizing Chronicle's memory feature for…
AI Research
93

Codex 'Chronicle' Research Preview Adds Memory for Daily Developer Context

A research preview of 'Chronicle' for Codex has been released. It enables the AI coding assistant to accumulate memories from a developer's daily workflow to improve context.

x.com/Apr 20, 2026/3 min read/Multi-Source
researchai assistantsopenai
A glowing quantum processor chip with intricate circuitry sits under blue light, surrounded by floating digital…
AI Research
95

Quantum Breakthrough: 100,000 Qubits Now Threatens Encryption

The estimated qubits required to break RSA encryption has collapsed from 1 billion in 2012 to just 10,000 in 2026, based on recent papers from Caltech, Google, and quantum startup Oratomic.

x.com/Apr 20, 2026/3 min read
ai-securitysecurityresearch
A glowing circuit board with a skull symbol embedded in the central processor, surrounded by scattered document…
AI Research
95

PoisonedRAG Attack Hijacks LLM Answers 97% of Time with 5 Documents

Researchers demonstrated that inserting only 5 poisoned documents into a 2.6 million document database can hijack a RAG system's answers 97% of the time, exposing critical vulnerabilities in 'hallucination-free' retrieval systems.

x.com/Apr 20, 2026/3 min read
ai securityllmsvulnerability
Two microchips on a circuit board with red glowing connections, symbolizing critical neural network parameters
AI Research
95

DNL Method Finds 2 Bits That Crash ResNet-50, Qwen3-30B

Researchers introduced Deep Neural Lesion (DNL), a method to find critical parameters. Flipping just two sign bits reduced ResNet-50 accuracy by 99.8% and Qwen3-30B reasoning to 0%.

x.com/Apr 20, 2026/3 min read
hardwareai securityvulnerability
A diagram comparing JEPA and LeWorldModel architectures, showing LeWorldModel avoiding representation collapse with…
AI Research
95

LeWorldModel Solves JEPA Collapse with 15M Params, Trains on Single GPU

Researchers published LeWorldModel, solving the representation collapse problem in Yann LeCun's JEPA architecture. The 15M-parameter model trains on a single GPU and demonstrates intrinsic physics understanding.

x.com/Apr 20, 2026/3 min read
world-modelsresearchcomputer-vision
A diagram showing an LLM processing a query, with a probe detecting uncertainty in hidden states, triggering a…
AI Research
85

Skill-RAG Uses Hidden-State Probes to Trigger Retrieval Only When Needed

Researchers introduced Skill-RAG, a system that uses hidden-state probing to detect when an LLM is about to fail, triggering targeted retrieval. This improves over uniform RAG baselines on HotpotQA, Natural Questions, and TriviaQA.

x.com/Apr 20, 2026/3 min read
nlpefficiencyagents
Person typing health question into smartphone with glowing chatbot interface and medical cross icon in background
AI Research
85

BBC Reports AI Chatbots Are Primary Health Advice Entry Point

The BBC reports AI chatbots have become a major front door for health advice. New evidence indicates hybrid human-AI systems outperform pure AI models in healthcare contexts.

x.com/Apr 20, 2026/3 min read
ai ethicspublic policyhealthcare
A live cockroach fitted with a tiny electronic backpack and sensor array, crawling along rubble in a test…
AI Research
85

NATO Tests SWARM Biotactics' AI-Guided Cyborg Cockroaches for Recon

NATO is evaluating a biohybrid system from German defense startup SWARM Biotactics, which uses AI to guide live cockroaches fitted with sensor backpacks through complex environments for military reconnaissance.

x.com/Apr 20, 2026/3 min read
roboticssurveillancedefense ai
Alibaba researchers' DCW method diagram showing wavelet-based SNR-t correction for diffusion models like FLUX and…
AI Research
85

Alibaba's DCW Fixes SNR-t Bias in Diffusion Models, Boosts FLUX & EDM

Alibaba researchers developed DCW, a wavelet-based method to correct SNR-t misalignment in diffusion models. The fix improves performance for models like FLUX and EDM with minimal computational cost.

x.com/Apr 20, 2026/3 min read
computer visionresearchgenerative ai
A diverse group of professionals in a modern office collaborate on laptops and tablets, with an AI interface visible…
AI Research
95

Gallup: 50% of US Workers Now Use AI on the Job, Doubling Since 2023

A Gallup survey of nearly 24,000 US workers in Q1 2026 shows 50% now use AI at work, up from just 21% in 2023. This marks a critical mass for enterprise AI tools and signals a shift from experimentation to operational integration.

x.com/Apr 20, 2026/3 min read
trendsresearchbusiness
Diagram showing a split LLM inference pipeline with a dedicated prefill server on the left and a separate decoding…
AI Research
85

Prefill-as-a-Service Paper Claims to Decouple LLM Inference Bottleneck

A research paper proposes a 'Prefill-as-a-Service' architecture to separate the heavy prefill computation from the lighter decoding phase in LLM inference. This could enable new deployment models where resource-constrained devices handle only the decoding step.

x.com/Apr 20, 2026/3 min read
edge computingresearchinference
A hooded figure sits before a glowing monitor displaying fragmented user profiles and connecting lines, symbolizing…
AI Research
85

LLMs Can De-Anonymize Users from Public Data, Study Warns

Large Language Models can now piece together a person's identity from their public online trail, rendering pseudonyms ineffective. This raises significant privacy and security concerns for internet users.

x.com/Apr 20, 2026/3 min read
privacyai ethicssecurity
ByteDance researcher presenting PersonaVLM diagram on screen, showing MLLM personalization improvement metrics and…
AI Research
97

ByteDance's PersonaVLM Boosts MLLM Personalization by 22.4%, Beats GPT-4o

ByteDance researchers unveiled PersonaVLM, a framework that transforms multimodal LLMs into personalized assistants with memory. It improves baseline performance by 22.4% and surpasses GPT-4o by 5.2% on personalized benchmarks.

x.com/Apr 20, 2026/3 min read
multimodal-aiagentsresearch
Satellite image showing varied terrain with buildings, roads, and vegetation, illustrating geospatial AI…
AI Research
88

OVRSISBenchV2: New 170K-Image Benchmark for Realistic Remote Sensing AI

A new benchmark, OVRSISBenchV2, with 170K images and 128 categories, sets a more realistic test for geospatial AI segmentation. The accompanying Pi-Seg model uses learnable semantic noise to broaden feature space and improve transfer.

arxiv.org/Apr 20, 2026/3 min read/Multi-Source
geospatialresearchcomputer-vision
A grid of colorful agent icons on a dark background, resembling a social deduction game interface, with one icon…
AI Research
100

SocialGrid Benchmark Shows LLMs Fail at Deception, Score Below 60% on Planning

Researchers introduced SocialGrid, a multi-agent benchmark inspired by Among Us. It shows state-of-the-art LLMs fail at deception detection and task planning, scoring below 60% accuracy.

arxiv.org/Apr 20, 2026/3 min read/Widely Reported
researchai agentsbenchmarks
A physics researcher studies equations on a whiteboard while a laptop displays a data graph, with scientific papers…
AI Research
100

PRL-Bench: LLMs Score Below 50% on End-to-End Physics Research Tasks

Researchers introduced PRL-Bench, a benchmark built from 100 recent Physical Review Letters papers, testing LLMs on end-to-end physics research. Top models scored below 50%, exposing a significant capability gap for autonomous scientific discovery.

arxiv.org/Apr 20, 2026/3 min read/Widely Reported
researchmachine learningai agents
A researcher analyzes a KWBench dashboard displaying LLM performance metrics across 223 game-theoretic tasks, with…
AI Research
100

KWBench: New Benchmark Tests LLMs' Unprompted Problem Recognition

Researchers introduced KWBench, a 223-task benchmark measuring if LLMs can recognize the governing game-theoretic problem in professional scenarios without being told what to look for. The best-performing model passed only 27.9% of tasks, highlighting a critical gap between task execution and situational understanding.

arxiv.org/Apr 20, 2026/3 min read/Widely Reported
researchai agentsbenchmarks
Diagram comparing teacher and student AI models, showing unsafe behaviors like deletion biases transferring through…
AI Research
100

Subliminal Transfer Study Shows AI Agents Inherit Unsafe Behaviors Despite

New research demonstrates unsafe behavioral traits in AI agents can transfer subliminally through model distillation, with students inheriting deletion biases despite rigorous keyword filtering. This exposes a critical security flaw in agent training pipelines.

arxiv.org/Apr 20, 2026/3 min read/Widely Reported
ai safetysecurityresearch
A glowing digital network of interconnected nodes and lines representing AI agents collaborating on a complex…
AI Research
85

Researchers Achieve Ultra-Long-Horizon Agentic Science with Cohesive AI Agents

A research team has developed AI agents capable of executing and maintaining coherent, long-horizon scientific research workflows. This addresses a core challenge in creating autonomous systems for complex discovery.

x.com/Apr 20, 2026/3 min read
agentsautonomyresearch
A futuristic digital network of interconnected glowing nodes and data streams, symbolizing AI agents autonomously…
AI Research
85

New Protocol Enables Self-Improving AI Agents with Auditable Lineage

Researchers have proposed a formal protocol for creating self-improving AI agent systems. The framework enables agents to autonomously evaluate and implement upgrades while maintaining auditable lineage and safe rollback options.

x.com/Apr 19, 2026/3 min read
ai safetyai agentsautonomous systems

Omar Sarayra Builds LLM Artifact Generator for AI Knowled…

AI Research
87

Omar Sarayra Builds LLM Artifact Generator for AI Knowledge Discovery

Omar Sarayra created a system that transforms dense LLM knowledge bases into consumable visual artifacts, like a pulse on HN AI discussions. He argues this format could become a new medium for staying current.

x.com/Apr 19, 2026/3 min read
human-computer interactionai agentsprototype
Fei-Fei Li gestures while explaining AI challenges, with a robotic arm and a cabinet with a vase on top visible in…
AI Research
85

Fei-Fei Li Explains Why 'Open the Top Drawer' Is a Hard AI Problem

AI pioneer Fei-Fei Li breaks down why a simple instruction like 'open the top drawer and watch out for the vase' represents a major unsolved challenge in robotics, requiring robust perception, commonsense reasoning, and efficient learning from sparse rewards.

x.com/Apr 19, 2026/3 min read
roboticscomputer visionmachine learning