Sakana AI developed Error Diffusion, a training method that adheres to Dale's principle and avoids backpropagation's weight transport requirement, achieving 96.7% accuracy on MNIST and 61.7% on CIFAR-…
OpenAI developed GPT-Red, an automated red-teaming model that outperformed human red-teamers 84% to 13% on prompt injection attacks, discovering a novel 'Fake Chain-of-Thought' attack class and reduci…
A developer guide demonstrates building a restaurant booking phone agent using the Patter SDK, featuring dynamic variables, guardrails, latency dashboards, and eval checks. The workflow integrates Twi…
MarkTechPost published a tutorial on building a Gin Config controlled PyTorch pipeline that separates training code from experiment configuration using .gin files. The pipeline includes configurable M…
PrismML released Bonsai 27B, a 27-billion parameter AI model compressed to 3.9GB via native 1-bit training, enabling on-device inference on an iPhone 17 Pro at 11 tokens per second. The model, availab…
A tutorial reconstructs the VideoAgent multi-agent pipeline for video editing, integrating intent parsing, graph planning, tool routing, and textual-gradient optimization with FFmpeg, Whisper, and oth…
MarkTechPost published a tutorial on building an autonomous data science agent using DeepAnalyze-8B, sandboxed code execution, and iterative analysis. The agent runs on a T4 GPU in Colab, processes e-…
Google Research, Google DeepMind, and university collaborators introduced SensorFM, a wearable health foundation model pretrained on over one trillion minutes of unlabeled sensor data from 5 million p…
Robbyant, Ant Group's embodied-intelligence unit, released LingBot-World-Infinity, a 14B causal video generation model that functions as an interactive world simulator. The model uses a Mixture of Bid…
A tutorial demonstrates building an autonomous AI co-scientist for EGFR C797S inhibitor discovery using ChEMBL, RDKit, SHAP, and BRICS. The system mines IC50 data, trains a scaffold-split Random Fores…
Synthetic Sciences released OpenScience, an open-source AI workbench for scientific research under the Apache-2.0 license. The platform supports any frontier or open-weight model via API keys, integra…
MarkTechPost published a tutorial on training Gemma-3 for structured mathematical reasoning using Tunix GRPO, LoRA adapters, and GSM8K rewards. The workflow includes environment setup, prompt formatti…
A new tutorial demonstrates building a multimodal retrieval pipeline called RAG-Anything that handles text, tables, equations, and images in Google Colab. The workflow uses OpenAI APIs to test naive, …
Anthropic launched Claude Sonnet 5 on June 30 with a two-month introductory pricing of $2 per million input tokens and $10 per million output tokens, set to increase 50% on September 1 to $3/$15. The …
EverMind has open-sourced EverOS, a local-first memory runtime for AI agents that stores memory as plain Markdown indexed by SQLite and LanceDB, featuring hybrid BM25 + vector retrieval, multimodal in…
DeepSeek open-sourced DSpark, a speculative decoding framework that accelerates per-user generation for DeepSeek-V4 by 57–85% over the MTP-1 baseline. The framework pairs a parallel draft backbone wit…
NVIDIA released Open-SWE-Traces, a dataset of agentic software-engineering trajectories for fine-tuning AI models. The tutorial processes the data from Hugging Face, normalizing conversations, parsing…
A tutorial demonstrates building a lightweight AI agent in Google Colab inspired by nanobot architecture, covering tool calling, session memory, skills, and MCP servers. The agent is built from scratc…
Gradium launched two real-time speech translation models, stt-translate and s2s-translate, covering 20 language pairs across English, French, German, Spanish, and Portuguese. The models collapse the s…
A tutorial demonstrates building an OpenHarness-style agent runtime from scratch, covering tool use, typed schemas, permissions, memory, skills, multi-agent coordination, and other core components. Th…