Engineering Insights

Scaling up BERT-like model Inference on modern CPU - Part 2

Introducing PlanetScale Managed on AWS and GCP

When the PlanetScale beta was released in May of 2021, we initially set out to launch on a shared tenancy model where databases from different organizations run in separate processes while sharing compute resources. Shortly after our beta launch, we saw high demand from organizations with regulatory requirements and compliance constraints that prohibited them from running in a shared tenancy model. It was clear that we needed to offer a single-tenancy model as well. For these customers who requi

Solving math word problems

We’ve trained a system that solves grade school math problems with nearly twice the accuracy of a fine-tuned GPT-3 model. It solves about 90% as many problems as real kids: a small sample of 9-12 year olds scored 60% on a test from our dataset, while our system scored 55% on those same problems.

New PlanetScale pricing: Scaler plan upgrades and our new enterprise plan

In May of this year, we launched PlanetScale, the database for developers, to easily use developer workflows on top of Vitess, the database technology that powers hyperscalers. From the outset, a goal of ours has been to make pricing transparent and friendly to users. We never want you to be surprised by a bill. The initial reception to this structure has been fantastic, with over 8,000 databases created in our free tier since May. We are dedicated to always refining and making things even simpl

Node.js 16.13.0 (LTS)

Next.js 12

Next.js 12 introduces a brand-new Rust compiler, Middleware (beta), React 18 Support, Native ESM Support, URL Imports, React Server Components (alpha), and more!

Course Launch Community Event

Large Language Models: A New Moore's Law?