engblogs

Explores the advantages, challenges, and applications of synthetic data in AI, emphasizing privacy preservation, data augmentation, and the importance of validation metrics to ensure model performance and mitigate bias.

MIT researchers developed FlowER, a generative AI model integrating physical constraints like mass and electron conservation to improve the accuracy of chemical reaction predictions using bond-electron matrix representations and large-scale experimental data.

OpenAI is enhancing ChatGPT to better support mental health and well-being through expert partnerships, advanced reasoning models, expanded crisis interventions, and strengthened teen protections including parental controls in a 120-day initiative.

Vijaye Raji joins OpenAI as CTO of Applications following the acquisition of Statsig, aiming to enhance AI-driven product engineering, experimentation, and reliable scalable systems for ChatGPT and beyond.

An in-depth exploration of rubric design for evaluating generative AI outputs, detailing types, measurement granularity, specificity, process vs. outcome evaluation, and methods including human, LLM, code-based, and reward model assessments.

Exploring the transformative impact of machine learning on biology and medicine, focusing on causal inference, multi-omics data integration, and innovative predictive models driving the next era of biomedical research.

OpenAI launches a $50M People-First AI Fund to support U.S.-based nonprofits applying innovative AI solutions for community impact and public good.

The Realtime API is now generally available featuring the advanced gpt-realtime speech-to-speech model with enhanced natural language understanding, low latency, multimodal inputs, SIP calling, and improved function calling for building production-ready voice agents.

MIT researchers have developed VaxSeer, an AI-driven tool leveraging protein language models and deep learning to accurately predict dominant influenza strains and optimize flu vaccine strain selection months in advance.

An introductory exploration of writing and optimizing matrix multiplication GPU kernels on NVIDIA's Blackwell architecture using Mojo, highlighting architectural advancements and programming paradigms critical for high-performance LLM workloads.

OpenAI and Anthropic conducted a collaborative safety evaluation of their AI models, comparing performance across instruction hierarchy, jailbreak resistance, hallucination rates, and scheming behavior to advance transparency and improve alignment techniques.

OpenAI presents a collective alignment approach using global public input to refine and update its Model Spec, ensuring AI behavior reflects diverse human values and priorities through a combination of automated and human-driven processes.