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
AI researcher pointing at a diagram of a feedback loop between a human conversation and a language model training…
AI Research
85

OpenClaw-RL Trains AI Agents on Conversation Feedback Without Manual Labels

OpenClaw-RL trains AI agents on natural conversation feedback, removing manual labeling. Uses evaluative and directive signals for continuous learning.

x.com/May 6, 2026/3 min read
ai-agentslanguage-modelsreinforcement-learning
MNEMA five-layer architecture stack: Substrate, Witness, Lattice, Protocol, Audit — each layer ablatable for falsifiability
AI ResearchBreakthrough
100

MNEMA: A Witness Lattice for Multi-Agent AI Memory

Today's agentic AI fails three ways: agents miscoordinate, memory gets quietly poisoned, and decisions can't be audited. A new EUMAS 2026 submission argues the fix is to stop treating memory as static records. Make it *living* — every memory unit becomes an autonomous cryptographic witness that interacts with other witnesses (agree, disagree, give birth to new witnesses, split, coalesce, retire), and decisions emerge from a fixed signed protocol rather than from a single orchestrator.

gentic_lab, devto_claudecode, hn_claude_code, arxiv_ir/May 6, 2026/3 min read/Widely Reported
anthropicai safetymulti-agent systems
NVIDIA and Unsloth engineers collaborate on a laptop, with code and performance graphs on screen showing a 25%…
AI Research
87

Unsloth × NVIDIA Cut LLM Fine-Tuning ~25% — Three Glue-Code Wins on Blackwell

Daniel & Michael Han at Unsloth, in collaboration with NVIDIA, published a joint guide quantifying three glue-code optimizations that combine for ~25% faster LLM training on B200 Blackwell hardware. The wins target overhead around the main kernels — caching packed-sequence metadata, double-buffered gradient checkpoint reloads, and a cheaper GPT-OSS MoE router using argsort + bincount. All three are merged via public PRs.

x.com/May 6, 2026/3 min read
ml systemsunslothfine-tuning
Researchers analyze a graph showing failure rates for long-horizon AI agents across extended tasks, with a laptop…
AI Research
75

Microsoft Paper Probes Long-Horizon Agent Generalization Gap

Microsoft Research paper on long-horizon agent generalization identifies failure modes and proposes improvements for extended tasks.

x.com/May 6, 2026/3 min read
agentsresearchgeneralization
A person typing on a laptop with code visible on the screen, surrounded by abstract security icons like locks and…
AI Research
100

Skills as Untrusted Code: A Security Precedent for Agent Runtimes

Paper argues agent skills are untrusted code until verified; runtimes must enforce verification gates to prevent supply-chain attacks, echoing decades of software security lessons.

x.com/May 5, 2026/3 min read/Widely Reported
ai securityresearchsupply chain
Two robotic arms with articulated hands at a tabletop, surrounded by tools and objects, illustrating bimanual…
AI Research
95

AllenAI's MolmoAct2: 720-Hour Bimanual Dataset, Beats GPT-5 on Robotics

AllenAI released MolmoAct2, an open robotics model with a 720-hour bimanual dataset, beating GPT-5 and Gemini Robotics on success rate (89.4% vs 82.1%) with 40% lower latency.

x.com/May 5, 2026/3 min read
open-sourceroboticsbenchmarks
A data scientist at a computer terminal displays a diagram comparing traditional vector database architecture to a…
AI Research
89

New RAG method ditches vector DB, threatens industry

New RAG method ditches vector DB, threatening incumbents. Claim from single tweet, no verification yet.

x.com/May 5, 2026/3 min read
vector databasesstartupsai
A stylized brain network with glowing nodes and connecting lines, representing AI self-play skill discovery without…
AI Research
85

Ctx2Skill: Self-Play Framework Lets LMs Discover Skills Without Labels

Ctx2Skill discovers skills from context via multi-agent self-play without labels. Outputs plug into any LM, targeting manual prompt engineering bottlenecks.

x.com/May 5, 2026/3 min read
skill discoveryself-playcontext learning
A soldier in camouflage reviews a laptop displaying a large language model interface with red warning indicators…
AI Research
92

ARMOR 2025: Military Safety Benchmark Exposes LLM Gaps Across 21 Models

ARMOR 2025 benchmark tests 21 LLMs against military legal doctrines, revealing critical safety gaps that civilian benchmarks miss.

arxiv.org/May 5, 2026/3 min read/Widely Reported
llm evaluationai safetydefense ai
Researchers presenting a fragmented face swapping benchmark with multiple face images and evaluation metrics on a…
AI Research
74

New CASIA Benchmark Exposes Fragmented Face Swapping Evaluation

CASIA researchers released a face swapping survey and benchmark on April 27, 2026, aiming to standardize evaluation across fragmented GAN and diffusion model methods.

arxiv.org/May 5, 2026/3 min read/Multi-Source
generative modelscomputer visionbenchmark
Two students at a hackathon table attach a sensor-laden wearable device to a person's arm, with a laptop displaying…
AI Research
77

MIT Hackathon Team Builds Wearable AI for Physical Movement Guidance

MIT hackathon team builds wearable AI for real-time physical movement guidance via sensors and on-device inference, demoed by @kimmonismus.

x.com/May 5, 2026/3 min read
wearablesaimit

Sakana AI 7B Conductor Hits SOTA on GPQA-Diamond via Orch…

AI Research
85

Sakana AI 7B Conductor Hits SOTA on GPQA-Diamond via Orchestration

Sakana AI's 7B Conductor model achieves SOTA on GPQA-Diamond and LiveCodeBench via orchestration of specialized sub-models, accepted at ICLR 2026.

x.com/May 5, 2026/3 min read
researchbenchmarksai models
ByteDance researchers present GenLIP framework diagram showing Vision Transformer processing images to directly…
AI Research
85

ByteDance GenLIP: ViT Predicts Language Tokens Directly with 8B Samples

ByteDance's GenLIP trains ViTs to predict language tokens directly with a single autoregressive objective, outperforming baselines on 8B samples.

x.com/May 4, 2026/3 min read
genlipvision transformerbytedance
Researcher presenting UniVidX interface on a large screen, showing video frames with RGB, depth, and alpha channels…
AI Research
85

UniVidX Generates Video From 1,000 Samples, SIGGRAPH 2026

UniVidX generates omni-directional video from <1,000 training samples, using diffusion priors with stochastic masking, accepted at SIGGRAPH 2026.

x.com/May 4, 2026/3 min read
diffusion-modelsvideo-generationcomputer-graphics
A network of interconnected glowing nodes representing multi-agent systems, with a central hub labeled Eywa linking…
AI Research
89

Recursive Multi-Agent Systems Top Hugging Papers; Eywa Bridges LLMs and Scientific Models

Recursive Multi-Agent Systems leads Hugging Papers with 242 upvotes. Eywa and OneManCompany signal a move from chat-based to structural agent collaboration.

x.com/May 3, 2026/3 min read
multi-agent systemshugging papersscientific ai

Claude Opus 4.7 Builds AlphaZero-Style Self-Play on Consu…

AI Research
100

Claude Opus 4.7 Builds AlphaZero-Style Self-Play on Consumer Hardware

Claude Opus 4.7 built AlphaZero self-play from scratch on consumer hardware in three hours, showing autonomous algorithmic code generation.

x.com/May 3, 2026/3 min read/Widely Reported
code generationanthropicreinforcement learning
A single 24GB GPU card with a 32B model fine-tuning diagram, showing RoundPipe's efficient memory usage for full…
AI Research
85

RoundPipe: Full Fine-Tune 32B Models on a Single 24GB GPU

RoundPipe fine-tunes 32B models on a single 24GB GPU with 1.5-2.2× speedups via round-robin pipeline dispatch.

x.com/May 3, 2026/3 min read
hugging-facefine-tuningpipeline-parallelism
Three robot-like figures with glowing blue screens for faces stand around a table, appearing to fail at a simple…
AI Research
85

Study: AI Agent Groups Fail at Simple Coordination Tasks

A cited study shows AI agent groups fail at simple coordination, challenging multi-agent system assumptions. No paper details disclosed.

x.com/May 2, 2026/3 min read
multi-agent systemsreliabilityai research
OpenAI's o1 model displayed on a screen next to a doctor's stethoscope and medical charts, symbolizing AI…
AI Research
87

o1 Outperforms Human Doctors on Medical Benchmarks & ER Cases

o1 beat human physicians on medical benchmarks and real ER cases, per a new paper. Authors urge prospective trials.

x.com/May 1, 2026/3 min read
reasoning modelshealthcare aiai research
Two humanoid robots with glowing blue eyes face each other across a digital stock-market display showing rising…
AI Research
100

Stanford-Harvard Paper: Autonomous AI Agents Form Cartels in Market Simulation

Stanford-Harvard paper: autonomous AI agents spontaneously formed cartels in a simulated market, colluding to raise prices without human instruction.

x.com/May 1, 2026/3 min read/Widely Reported
economicsai safetyautonomous agents
A diagram shows a multi-step RAG process where a user query triggers iterative retrieval and reasoning loops, with…
AI Research
75

RAG's New Frontier: When to Retrieve During Reasoning

A new RAG paradigm retrieves at multiple reasoning steps via a learned gate, boosting multi-hop QA by 15-20% on HotpotQA.

x.com/May 1, 2026/3 min read
reasoningllmsresearch
A scientist examines a genomic sequence on a computer monitor while an AI assistant highlights a previously…
AI Research
85

Claude Solves Bioinformatics Problems Human Experts Miss

Anthropic shows Claude solves 23 bioinformatics problems human experts missed, catching errors in genomic analyses.

x.com/May 1, 2026/3 min read
claudeanthropicscientific ai
Mexican woman smiling while using smartphone with mental health app, surrounded by warm home setting
AI Research
85

AI Chatbot Improves Mexican Women's Mental Health by 0.3 SD in RCT

AI therapy chatbot RCT on Mexican women: 0.3 SD mental health improvement over 6 months, no severe case increase, plus labor market gains.

x.com/May 1, 2026/3 min read
rctmental healthdigital health
Screenshot of Qwen-Scope interface showing 81k feature activations across 64 layers in Qwen3.5-27B, with a neural…
AI Research
87

Qwen3.5-27B Gets Sparse Autoencoders: 81k Features Exposed

Qwen released Qwen-Scope, adding Sparse Autoencoders to Qwen3.5-27B, exposing 81k features across 64 layers for steerable inference.

x.com/Apr 30, 2026/3 min read
open sourceaiinterpretability
A Microsoft researcher points to a screen displaying a corrupted document with red error highlights alongside a…
AI Research
90

Microsoft: LLMs Corrupt 25% of Docs in Long Edits

Microsoft paper shows LLMs corrupt ~25% of documents across 52 domains during 20-edit sessions, with failures compounding silently.

x.com/Apr 30, 2026/3 min read
llm reliabilityagentic systemsdocument ai
A neural network diagram with some pathways dimmed or collapsed, illustrating reduced neural activity
AI Research
87

LLMs Shrink Neural Activity When Confused, New Paper Shows

LLMs compress neural activity when confused, measurable as a sparsity signal. Paper 2603.03415 proposes using this for adaptive prompting.

x.com/Apr 29, 2026/3 min read
uncertaintyllmsresearch
Engineers collaborating around a terminal screen displaying code and test results, with a diagram of agentic harness…
AI Research
100

Agentic Harness Engineering Boosts Coding Agents 7% on Terminal-Bench 2

Agentic Harness Engineering introduces a structured approach to evolving coding-agent harnesses, using revertible components, condensed experience, and falsifiable decisions. On Terminal-Bench 2, pass@1 climbs from 69.7% to 77.0% in ten iterations, beating human-designed baselines.

x.com/Apr 29, 2026/3 min read/Multi-Source
coding agentsagentic systemsharness engineering
Bar chart comparing vector database and graph retrieval performance on multi-hop reasoning and summarization tasks…
AI Research
75

Vector DBs Can't Reason: GraphRAG-Bench Shows 83.6% Gap on Complex Queries

FalkorDB's GraphRAG-Bench benchmarks show vector databases struggle on multi-hop reasoning (83.6% gap) and contextual summarization (85.1% gap), highlighting graph-based retrieval's advantage for complex queries.

x.com/Apr 29, 2026/3 min read
vector databasesai benchmarksfalkordb

GPT-5.5 Pro Leapfrogs on Epoch Benchmark; Base Model Beat…

AI Research
99

GPT-5.5 Pro Leapfrogs on Epoch Benchmark; Base Model Beats Prior Pro

A tweet from @kimmonismus reveals GPT-5.5 Pro shows significant Epoch benchmark gains, and the non-Pro GPT-5.5 surpasses GPT-5.4 Pro, suggesting major efficiency improvements at OpenAI.

x.com/Apr 29, 2026/3 min read/Multi-Source
gpt-5.5benchmarksai efficiency
Xiaomi MiMo 2.5 Pro logo beside Arena leaderboard chart, with MIT open-source badge and Opus 4.5 ranked lower
AI Research
97

Xiaomi MiMo 2.5 Pro Beats Opus 4.5 on Arena, MIT License

Xiaomi's MiMo v2.5 Pro, an open-source model under MIT license, has achieved a higher Arena score than Opus 4.5, signaling a major shift in competitive AI performance.

x.com/Apr 29, 2026/3 min read/Multi-Source
open sourcexiaomibenchmarks