Goblin
News
AI news, distilled.
|
News
About
News
About
Titles
Summaries
7
GPT-6 Astra Cracks Previously Unsolved WWI German Military Cipher
Updated Sep 19
Research
2
1h ago
7
GPT-6 Astra Cracks Previously Unsolved WWI German Military Cipher
Upd Sep 19
Research
· 2 srcs · 1h ago
Discuss
7
Z.ai's GLM-5.3 Builds Its Own Production Inference Stack in Early Recursive Self-Improvement Step
Research
2
16h ago
7
Z.ai's GLM-5.3 Builds Its Own Production Inference Stack in Early Recursive Self-Improvement Step
Research
· 2 srcs · 16h ago
Discuss
8
Claude Achieves 4x Speedup on 30+ Biomolecular Models and Launches $1M Protein Design Competition
Research
1
1d ago
8
Claude Achieves 4x Speedup on 30+ Biomolecular Models and Launches $1M Protein Design Competition
Top
Research
· 1 src · 1d ago
Discuss
8
Anthropic Quietly Sets Up Biology Wet Lab for AI-Driven Drug Discovery
Updated Sep 19
Research
2
16h ago
8
Anthropic Quietly Sets Up Biology Wet Lab for AI-Driven Drug Discovery
Top
Upd Sep 19
Research
· 2 srcs · 16h ago
Discuss
6
Amateur Researcher Claims AI-Assisted Lean Proof of Conway's 50-Year-Old Refinement Conjecture
Research
1
1d ago
6
Amateur Researcher Claims AI-Assisted Lean Proof of Conway's 50-Year-Old Refinement Conjecture
Research
· 1 src · 1d ago
Discuss
6
NVIDIA Researchers Release Agora: Git-Backed Shared Memory for Collaborative AI Research Agents
Research
1
1d ago
6
NVIDIA Researchers Release Agora: Git-Backed Shared Memory for Collaborative AI Research Agents
Research
· 1 src · 1d ago
Discuss
6
Sakana AI Forms Frontier Intelligence Group to Research Post-Transformer AI Paths
Research
1
1d ago
6
Sakana AI Forms Frontier Intelligence Group to Research Post-Transformer AI Paths
Research
· 1 src · 1d ago
Discuss
8
OpenAI Staff Expect to Solve Hodge Conjecture After Navier-Stokes Breakthrough
Research
1
2d ago
8
OpenAI Staff Expect to Solve Hodge Conjecture After Navier-Stokes Breakthrough
Top
Research
· 1 src · 2d ago
Discuss
7
HarnessTax Study: Coding Agent Harness Choice Creates Up to 5x Cost Gap With Minimal Impact on Task Success
Research
1
2d ago
7
HarnessTax Study: Coding Agent Harness Choice Creates Up to 5x Cost Gap With Minimal Impact on Task Success
Research
· 1 src · 2d ago
Discuss
7
Stanford Lab Publishes Science Paper on AI Agent 'Co-Scientists' for Accelerated Drug Discovery
Research
1
2d ago
7
Stanford Lab Publishes Science Paper on AI Agent 'Co-Scientists' for Accelerated Drug Discovery
Research
· 1 src · 2d ago
Discuss
7
Pew Research: Majorities in 34 of 37 Countries Fear AI Will Cause Net Job Losses
Updated Sep 18
Research
3
1d ago
7
Pew Research: Majorities in 34 of 37 Countries Fear AI Will Cause Net Job Losses
Upd Sep 18
Research
· 3 srcs · 1d ago
Discuss
6
Apple ML Research: Value Induction in LLM Post-Training Creates Sycophancy Risk and Unintended Behavioral Side Effects
Research
1
2d ago
6
Apple ML Research: Value Induction in LLM Post-Training Creates Sycophancy Risk and Unintended Behavioral Side Effects
Research
· 1 src · 2d ago
Discuss
6
JiT-DDT Architecture Trains Text-to-Image Models 3.6× Faster With 4× Higher Resolution
Research
1
2d ago
6
JiT-DDT Architecture Trains Text-to-Image Models 3.6× Faster With 4× Higher Resolution
Research
· 1 src · 2d ago
Discuss
6
Researchers Propose 'Infinite-Parameter LLM' Architecture That Learns from Live Interaction Data
Research
1
2d ago
6
Researchers Propose 'Infinite-Parameter LLM' Architecture That Learns from Live Interaction Data
Research
· 1 src · 2d ago
Discuss
6
Meta Introduces FLAT: Unified Continuous Token Representation for Multimodal AI
Research
1
2d ago
6
Meta Introduces FLAT: Unified Continuous Token Representation for Multimodal AI
Research
· 1 src · 2d ago
Discuss
7
Researchers Identify "Matthew Effect" in RL Training for LLMs, Propose "Never Give Up" Fix
Research
1
3d ago
7
Researchers Identify "Matthew Effect" in RL Training for LLMs, Propose "Never Give Up" Fix
Research
· 1 src · 3d ago
Discuss
7
OpenAI Researcher Noam Brown: Recursive Self-Improvement Is Company's Top Priority
Research
1
3d ago
7
OpenAI Researcher Noam Brown: Recursive Self-Improvement Is Company's Top Priority
Research
· 1 src · 3d ago
Discuss
7
Dream-RSI: Google DeepMind-Affiliated Researchers Propose Recursive Self-Improvement via Dreaming
Research
2
3d ago
7
Dream-RSI: Google DeepMind-Affiliated Researchers Propose Recursive Self-Improvement via Dreaming
Research
· 2 srcs · 3d ago
Discuss
7
Sakana AI Trains 1,000-Layer Neural Networks Without Backpropagation
Research
1
3d ago
7
Sakana AI Trains 1,000-Layer Neural Networks Without Backpropagation
Research
· 1 src · 3d ago
Discuss
7
AI Is Suppressing Wage Growth Across Millions of Jobs, New Research Finds
Research
1
3d ago
7
AI Is Suppressing Wage Growth Across Millions of Jobs, New Research Finds
Research
· 1 src · 3d ago
Discuss
7
University of Manchester Deploys NVIDIA Earth-2 AI for UK-Wide Air Pollution Forecasting
Research
1
3d ago
7
University of Manchester Deploys NVIDIA Earth-2 AI for UK-Wide Air Pollution Forecasting
Research
· 1 src · 3d ago
Discuss
6
OpenAI Research Finds Workers Increasingly Using AI for Tasks Outside Their Own Occupations
Research
1
3d ago
6
OpenAI Research Finds Workers Increasingly Using AI for Tasks Outside Their Own Occupations
Research
· 1 src · 3d ago
Discuss
6
Recursive AI System Discovers Architecture Programs Failure Evolution in Hierarchical Metamaterials
Research
1
3d ago
6
Recursive AI System Discovers Architecture Programs Failure Evolution in Hierarchical Metamaterials
Research
· 1 src · 3d ago
Discuss
6
Researchers Introduce Shared Selective Persistent Memory for Agentic LLM Systems
Research
1
3d ago
6
Researchers Introduce Shared Selective Persistent Memory for Agentic LLM Systems
Research
· 1 src · 3d ago
Discuss
6
BITCOS Paper Breaks 1.58-Bit-Per-Weight Barrier for Ternary LLM Storage
Research
1
3d ago
6
BITCOS Paper Breaks 1.58-Bit-Per-Weight Barrier for Ternary LLM Storage
Research
· 1 src · 3d ago
Discuss
7
StepAudio 3 Gen: Unified Discrete Autoregressive Model Achieves SOTA on TTS and Voice Design Across Seven Audio Types
Research
1
4d ago
7
StepAudio 3 Gen: Unified Discrete Autoregressive Model Achieves SOTA on TTS and Voice Design Across Seven Audio Types
Research
· 1 src · 4d ago
Discuss
7
Mozilla Report: Open Chinese AI Models Within 4.4 Months of US Frontier at 5× Lower Cost
Updated Sep 16
Research
2
3d ago
7
Mozilla Report: Open Chinese AI Models Within 4.4 Months of US Frontier at 5× Lower Cost
Upd Sep 16
Research
· 2 srcs · 3d ago
Discuss
7
Google Announces Wave of AI Science Breakthroughs Across Health, Genomics, and Weather
Research
1
4d ago
7
Google Announces Wave of AI Science Breakthroughs Across Health, Genomics, and Weather
Research
· 1 src · 4d ago
Discuss
7
Google Launches Interactive AI & Economy ATLAS Tool with Global Data on AI Adoption by Occupation and Region
Research
1
4d ago
7
Google Launches Interactive AI & Economy ATLAS Tool with Global Data on AI Adoption by Occupation and Region
Research
· 1 src · 4d ago
Discuss
7
MIT Report Warns AI Is Causing 'Cognitive Surrender' in Students, Urges Universities to Redesign Curricula
Research
1
4d ago
7
MIT Report Warns AI Is Causing 'Cognitive Surrender' in Students, Urges Universities to Redesign Curricula
Research
· 1 src · 4d ago
Discuss
7
OpenAI Foundation Launches 'Data for Public Health' Program with $40M+ in Initial Grants
Research
1
4d ago
7
OpenAI Foundation Launches 'Data for Public Health' Program with $40M+ in Initial Grants
Research
· 1 src · 4d ago
Discuss
7
ALTK-Evolve Consistency Analyzer Exposes Hidden AI Agent Reliability Gap
Research
1
3d ago
7
ALTK-Evolve Consistency Analyzer Exposes Hidden AI Agent Reliability Gap
Research
· 1 src · 3d ago
Discuss
6
ICML Research: LLM Judge Agreement May Reflect Shared Training Bias, Not True Reliability
Research
1
4d ago
6
ICML Research: LLM Judge Agreement May Reflect Shared Training Bias, Not True Reliability
Research
· 1 src · 4d ago
Discuss
6
Richard Socher Launches Recursive With $4.65B Seed Round to Build Self-Improving AI
Research
1
4d ago
6
Richard Socher Launches Recursive With $4.65B Seed Round to Build Self-Improving AI
Research
· 1 src · 4d ago
Discuss
7
ARC Prize Announces ARC-AGI-4 Benchmark Targeting Autonomous Open-Ended AI Innovation
Research
1
5d ago
7
ARC Prize Announces ARC-AGI-4 Benchmark Targeting Autonomous Open-Ended AI Innovation
Research
· 1 src · 5d ago
Discuss
7
Expert Re-Grading Reveals AI Physics Benchmarks Are Broken — Frontier Models Near-Saturating Once Fixed
Research
1
5d ago
7
Expert Re-Grading Reveals AI Physics Benchmarks Are Broken — Frontier Models Near-Saturating Once Fixed
Research
· 1 src · 5d ago
Discuss
7
Google ToolGrad Flips AI Tool-Use Training: Answer-First Data Generation Achieves 99.8% Success
Research
1
5d ago
7
Google ToolGrad Flips AI Tool-Use Training: Answer-First Data Generation Achieves 99.8% Success
Research
· 1 src · 5d ago
Discuss
6
Recurrent Looped Transformer: New Architecture Proposes Unbounded Temporal Depth for AI Reasoning
Updated Sep 15
Research
2
4d ago
6
Recurrent Looped Transformer: New Architecture Proposes Unbounded Temporal Depth for AI Reasoning
Upd Sep 15
Research
· 2 srcs · 4d ago
Discuss
6
MIT's HardFlow Algorithm Enables Generative AI to Satisfy Hard Constraints Without Retraining
Research
1
5d ago
6
MIT's HardFlow Algorithm Enables Generative AI to Satisfy Hard Constraints Without Retraining
Research
· 1 src · 5d ago
Discuss
7
Real-SWE Benchmark Tests AI Coding Agents on Private Enterprise Codebases
Research
1
6d ago
7
Real-SWE Benchmark Tests AI Coding Agents on Private Enterprise Codebases
Research
· 1 src · 6d ago
Discuss
7
NSF Launches $1.5 Billion 'X-Labs' Initiative for AI and Quantum Research
Research
1
6d ago
7
NSF Launches $1.5 Billion 'X-Labs' Initiative for AI and Quantum Research
Research
· 1 src · 6d ago
Discuss
7
Claude Fable 5.1 Autonomously Solves Two 370-Year-Old Historical Ciphers from Urquhart's Works
Research
1
6d ago
7
Claude Fable 5.1 Autonomously Solves Two 370-Year-Old Historical Ciphers from Urquhart's Works
Research
· 1 src · 6d ago
Discuss
8
25 Fields Medallists Including Terence Tao Sign Declaration Against AI Math Benchmarking
Updated Sep 16
Research
4
3d ago
8
25 Fields Medallists Including Terence Tao Sign Declaration Against AI Math Benchmarking
Top
Upd Sep 16
Research
· 4 srcs · 3d ago
Discuss
7
OpenAI Pulls Caltech Math Hackathon Sponsorship After "Slop Mathematics" Backlash
Updated Sep 11
Research
2
Sep 11
7
OpenAI Pulls Caltech Math Hackathon Sponsorship After "Slop Mathematics" Backlash
Upd Sep 11
Research
· 2 srcs · Sep 11
Discuss
7
Scaling Web-Video Pre-training Consistently Improves Real Robot Task Performance, Study Confirms
Research
1
Sep 11
7
Scaling Web-Video Pre-training Consistently Improves Real Robot Task Performance, Study Confirms
Research
· 1 src · Sep 11
Discuss
6
Meta Releases WearableQA: AI Health Reasoning Benchmark Built on Real Wearable Data
Research
1
Sep 11
6
Meta Releases WearableQA: AI Health Reasoning Benchmark Built on Real Wearable Data
Research
· 1 src · Sep 11
Discuss
6
Apple ML Research Introduces SimpleDesign for Joint Protein Sequence and Structure Co-design
Research
1
Sep 11
6
Apple ML Research Introduces SimpleDesign for Joint Protein Sequence and Structure Co-design
Research
· 1 src · Sep 11
Discuss
6
Salesforce Research: Imitation Fine-tuning Breaks Agent Harness Fit; On-Policy Correction Fixes It
Research
1
Sep 11
6
Salesforce Research: Imitation Fine-tuning Breaks Agent Harness Fit; On-Policy Correction Fixes It
Research
· 1 src · Sep 11
Discuss
7
Research Finds AI Reasoning Models Produce Fractal Patterns, Explaining "Overthinking"
Research
1
Sep 10
7
Research Finds AI Reasoning Models Produce Fractal Patterns, Explaining "Overthinking"
Research
· 1 src · Sep 10
Discuss
6
Perplexity Releases Q2D-Web: Large-Scale Web Retrieval Benchmark with 190M Documents
Research
1
Sep 10
6
Perplexity Releases Q2D-Web: Large-Scale Web Retrieval Benchmark with 190M Documents
Research
· 1 src · Sep 10
Discuss
Filters
Signal
Title
Category
Sources
Posted
Discuss
