plaid-ed76da2d·3 events·first seen Aliases: PLAID
OpenAI is rolling out a personal finance feature in ChatGPT for Pro users in the United States, allowing them to connect financial accounts via Plaid and view spending, bills, subscriptions, net worth, and investment data in a unified dashboard. Users can ask ChatGPT questions grounded in their actual financial context, though the system is explicitly limited to advisory functions and cannot execute transactions. The feature is available on web and iOS and is being gradually deployed.
This paper revisits continuous diffusion language models (DLMs) by introducing RePlaid, an updated version of Plaid that aligns its architecture with modern discrete DLMs. RePlaid establishes the first scaling law for continuous DLMs competitive with discrete approaches, achieving a compute gap of only 20× versus autoregressive models and a state-of-the-art perplexity bound of 22.1 on OpenWebText among continuous DLMs. The authors provide theoretical analysis showing that likelihood-based training naturally yields linear cross-entropy over time and creates structured embedding geometries, explaining the performance gains.
PLAID is a generative model that simultaneously produces protein 1D sequences and 3D all-atom structures by learning a diffusion model over the latent space of ESMFold, a protein folding model. It requires only sequence data for training—leveraging databases 2-4 orders of magnitude larger than structure databases—and decodes structure at inference via frozen folding model weights. The approach supports compositional prompting by function and organism, addressing practical drug-design constraints like humanization and solubility. A companion compression model, CHEAP, addresses the high-dimensionality of transformer latent spaces to make the diffusion training tractable.