epoch ai
27 articles about epoch ai in AI news
Epoch AI's CursorBench Benchmarks AI Code Editing at Scale
Epoch AI launched CursorBench, a 500-task benchmark for AI code editors. It reveals a 15% accuracy gap vs. humans and 3x latency variance.
SciCode: Epoch AI Launches Benchmark Measuring AI Research Ability
Epoch AI launched SciCode benchmark testing LLMs on real research coding tasks. Top models score below 30%, exposing gap between coding benchmarks and scientific ability.
Epoch AI: Hormuz LNG Shock Absorbed by Chip Margins, Gulf Investment is AI Risk
A new analysis from Epoch AI Research finds the Strait of Hormuz conflict's energy shock is manageable for AI infrastructure, but the real threat is the potential drying up of Gulf capital investment, crucial for projects like Stargate UAE.
AI Data Center Scale Doubles Every 7 Months, Epoch Finds
Epoch AI finds AI data center scale doubles every 7 months, driven by Google, Microsoft, and Amazon investments. This accelerates beyond the earlier 12-month cycle, raising training cost projections to $10 billion by 2028.
OSWorld 2.0 Launches, Tests AI Agents on 1,500 Desktop Tasks
Epoch AI released OSWorld 2.0 with 1,500 desktop tasks, up from 369 in v1, testing AI agents on adversarial and cross-application workflows.
Nvidia B200 Costs $6,400 to Produce, Gross Margin Hits 82%
Epoch AI estimates Nvidia's B200 GPU costs $5,700–$7,300 to produce, with HBM memory and advanced packaging accounting for two-thirds of the cost. At a $30k–$40k sale price, chip-level gross margins reach ~82%, though rack-scale margins may be lower.
Open-Weight Models Trail Frontier AI by Four Months: EpochAI
EpochAI finds open-weight models trail frontier closed-source models by four months, a small gap reflecting rapid catch-up.
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.
WiT: Waypoint Diffusion Transformers Achieve FID 2.09 on ImageNet 256×256 in 265 Epochs, Matching JiT-L/16 Efficiency
Researchers introduced WiT, a diffusion transformer that uses semantic waypoints from pretrained vision models to resolve trajectory conflicts in pixel-space flow matching. It matches the performance of JiT-L/16 at 600 epochs in just 265 epochs, achieving an FID of 2.09 on ImageNet 256×256.
Colossus 2: xAI's Memphis Cluster Hits 300,000 GPUs
xAI's Colossus 2 hits 300,000 GPUs, targeting 1M by year-end. Training Grok-3, the $6B cluster challenges OpenAI and Google.
SemiAnalysis: Pretraining Dead for All but Frontier Labs
@SemiAnalysis_ declares pretraining dead for non-frontier labs, citing 'Pretrainitis' as vanity-driven waste. Prompt engineering offers higher ROI.
Time's First AI A-List: Alibaba, ByteDance, Zhipu AI Make Cut
Time magazine named Alibaba, ByteDance, and Zhipu AI among its first AI-specific top 10 list, alongside six US companies and France's Mistral AI. The recognition highlights China's growing global influence through open-source models and consumer AI apps.
A Practical Guide to Fine-Tuning Open-Source LLMs for AI Agents
This Portuguese-language Medium article is Part 2 of a series on LLM engineering for AI agents. It provides a hands-on guide to fine-tuning an open-source model, building on a foundation of clean data and established baselines from Part 1.
VMLOps Launches 'Algorithm Explorer' for Real-Time Visualization of AI Training Dynamics
VMLOps released Algorithm Explorer, an interactive tool that visualizes ML training in real-time, showing gradients, weights, and decision boundaries. It combines math, visuals, and code to aid debugging and education.
China Surpasses US in AI Research Authorship with 2,152 First-Author Researchers in 2024
China now leads the US in first-author AI research contributions, with 2,152 researchers versus 1,810. This marks the first time China has overtaken the US in this key metric of research leadership.
Fine-Tuning LLMs While You Sleep: How Autoresearch and Red Hat Training Hub Outperformed the HINT3 Benchmark
Automated fine-tuning tools now let you run hundreds of training experiments overnight for under $50. Here's how Autoresearch and Red Hat's platform outperformed HINT3, and the tools you can use today.
Jensen Huang Predicts AI Training Shift to Synthetic Data, Compute as New Bottleneck
NVIDIA CEO Jensen Huang states AI training is moving from real-world to synthetic data, with compute power becoming the primary constraint as AI-generated data quality improves.
Goal-Driven Data Optimization: Training Multimodal AI with 95% Less Data
Researchers introduce GDO, a framework that optimizes multimodal instruction tuning by selecting high-utility training samples. It achieves faster convergence and higher accuracy using 5-7% of the data typically required. This addresses compute inefficiency in training vision-language models.
AI Now Surpasses Human Experts in Technical Domains, Study Finds
New research mapping AI capabilities to human expertise reveals frontier models have already surpassed domain experts in technical and scientific benchmarks. The study forecasts AI will reach top-performer human levels by late 2027.
The Great GPU Scramble: How Hardware Shortages Are Defining the AI Arms Race
Oracle founder Larry Ellison identifies GPU acquisition as the primary bottleneck in AI development, with companies racing to secure limited hardware for breakthroughs in medicine, video generation, and autonomous systems.
Beyond Better Models: The Compute Scaling Revolution Driving AI's Next Leap
New analysis reveals that scaling compute infrastructure may deliver 10× annual efficiency gains in AI development, surpassing algorithmic improvements alone. The real leverage comes from combining innovative ideas with massive computational resources.
Beyond Deterministic Benchmarks: How Proxy State Evaluation Could Revolutionize AI Agent Testing
Researchers propose a new LLM-driven simulation framework for evaluating multi-turn AI agents without costly deterministic backends. The proxy state-based approach achieves 90% human-LLM judge agreement while enabling scalable, verifiable reward signals for agent training.
ByteDance iLLaDA: 8B Diffusion LM Matches Qwen2.5 Base, Lags on Instruct
ByteDance iLLaDA, an 8B diffusion LM trained on 12T tokens, matches Qwen2.5 7B on base benchmarks (63.9 vs 63.3) but trails 10 points after instruction tuning, revealing the alignment gap for diffusion models.
Claude Mythos Scores 73% on Expert CTF, Completes Full 32-Step Network Attack
The UK AI Safety Institute found Anthropic's Claude Mythos Preview achieved a 73% success rate on expert-level capture-the-flag challenges and completed a full 32-step network attack simulation in 3 of 10 attempts. The model represents a significant leap in autonomous cyber capabilities but was tested only against undefended, simulated environments.
TensorFlow Playground Interactive Demo Updated for 2026, Enabling Real-Time Neural Network Visualization
The TensorFlow Playground, an educational web tool for visualizing neural networks, has been updated. Users can now adjust hyperparameters and watch the model train and visualize decision boundaries in real-time.
Developer Swaps Dash Cam Analysis for Gemma 4 & Falcon Perception
A developer announced they are replacing their entire dash cam video analysis system with Google's Gemma 4 and Falcon Perception models, signaling a practical shift towards newer, specialized multimodal models for real-time edge applications.
CoRe Framework Integrates Equivariant Contrastive Learning for Medical Image Registration, Surpassing Baseline Methods
Researchers propose CoRe, a medical image registration framework that jointly optimizes an equivariant contrastive learning objective with the registration task. The method learns deformation-invariant feature representations, improving performance on abdominal and thoracic registration tasks.