InfoQ Homepage News
-
Google Publishes LLM Self-Correction Algorithm SCoRe
Researchers at Google DeepMind recently published a paper on Self-Correction via Reinforcement Learning (SCoRe), a technique for improving LLMs' ability to self-correct when solving math or coding problems. Models fine-tuned with SCoRe achieve improved performance on several benchmarks compared to baseline models.
-
Uber Completes Major MySQL Fleet Upgrade, Boosting Performance and Security
Uber has detailed improvements to its MySQL database fleet by upgrading from version 5.7 to 8.0. The team wanted to take advantage of performance and concurrency improvements in newer versions of MySQL, and because MySQL 5.7 was reaching end-of-life in October 2023. The work took over a year and involved upgrading more than 2,100 clusters and 16,000 nodes across 19 production zones in 3 regions.
-
OpenAI Launches Public Beta of Realtime API for Low-Latency Speech Interactions
OpenAI launched the public beta of the Realtime API, offering developers the ability to create low-latency, multimodal voice interactions within their applications. Additionally, audio input/output is now available in the Chat Completions API, expanding options for voice-driven applications. Early feedback highlights limited voice options and response cutoffs.
-
Java News Roundup: JDK 24, Tomcat 11.0, Cassandra 5.0, EclipseStore 2.0, Payara Platform, Ktor 3.0
This week's Java roundup for October 7th, 2024, features news highlighting: JEP 489, Vector API (Ninth Incubator), and JEP 484, Class-File API, targeted for JDK 24; the release of Apache projects, Tomcat 11.0.0 and Cassandra 5.0.0; the release of EclipseStore 2.0.0; the October 2024 Payara Platform release; and the release of Ktor 3.0.0.
-
Ephemeral IDs: Cloudflare's Latest Tool for Fraud Detection
During its recent Birthday Week, Cloudflare introduced Ephemeral IDs, a new feature for fraud detection. The tool identifies fraudulent activity—whether from bots or humans—by linking behavior to a specific client rather than an IP address.
-
Scaling Uber’s Batch Data Platform: a Journey to the Cloud with Data Mesh Principles
Some months ago, Uber started the migration to the cloud, on Google Cloud Platform (GCP), of its batch data analytics and machine learning platform. In a recent post on its engineering blog, Uber provided additional information regarding its batch data cloud migration that incorporated crucial data mesh principles.
-
Maybe WebAssembly Is the Next Evolutionary Step From Containers: Fermyon at InfoQ DevSummit Munich
During her presentation at the inaugural edition of the InfoQ Dev Summit Munich, Danielle Lancashire, principal software engineer at Fermyon and co-chair of the CNCF wasm-wg, hinted at WebAssembly containers as a greener alternative and a potential evolution from the current containerised approach to serverless computing.
-
NVIDIA Unveils NVLM 1.0: Open-Source Multimodal LLM with Improved Text and Vision Capabilities
NVIDIA unveiled NVLM 1.0, an open-source multimodal large language model (LLM) that performs strongly on both vision-language and text-only tasks. NVLM 1.0 shows improvements in text-based tasks after multimodal training, standing out among current models. The model weights are now available on Hugging Face, with the training code set to be released shortly.
-
InfoQ Dev Summit Munich: How to Optimize Java for the 1BRC
Java applications passed the 1 Billion Row Challenge (1BRC) in 1.5 seconds. 1BRC creator Gunnar Morling detailed their optimizations at the InfoQ Dev Summit Munich 2024. General optimizations applicable to all Java applications cut the runtime from 290 seconds to 20 seconds. Getting to 1.5 seconds required niche optimizations that most Java applications should forego, except for possibly GraalVM.
-
OpenAI Developer Day 2024 (SF) Announces Real-Time API, Vision Fine-Tuning, and More
On October 1, 2024, OpenAI SF DevDay unveiled innovative features, including a Real-Time API enabling instant voice interactions and function calling. Enhanced model distillation and vision fine-tuning empower developers to customize AI for diverse applications. Upcoming events in London and Singapore will further expand these capabilities.
-
The Linux Kernel to Support Real-Time Scheduling out-of-the-Box
Linux 6.12 will officially include support for real-time processing in its mainline thanks to a PR that enables PREEMPT_RT on all supported architectures. While aimed at applications requiring deterministic time guarantees, like avionics, robotics, automotive, and communications, it could bring improvements to user experience on the desktop, too.
-
Setting up a Data Mesh Organization
A data mesh organization: producers, consumers, and the platform. According to Matthias Patzak, the mission of the platform team is to make the lives of the producer and consumers simple, efficient and stress free. Data must be discoverable and understandable, trustworthy, and shared securely and easily across the organization.
-
Hugging Face Upgrades Open LLM Leaderboard v2 for Enhanced AI Model Comparison
Hugging Face has recently released Open LLM Leaderboard v2, an upgraded version of their benchmarking platform for large language models. Hugging Face created the Open LLM Leaderboard to provide a standardized evaluation setup for reference models, ensuring reproducible and comparable results.
-
JFrog Integrates Runtime Security for Enhanced DevSecOps Platform
JFrog has introduced JFrog Runtime to its suite of security capabilities, adding real-time vulnerability detection to its software supply chain platform. This update is aimed at developers and DevSecOps teams working with Kubernetes clusters and cloud-native applications.
-
Data Teams Survey: Lag in DataOps and Value Delivered
We report on Jesse Anderson's 2024 Data Teams Survey which showed a lag in DataOps capabilities, slow LLM adoption, and a concerning decline in perceived value creation by data teams. It called out the importance of teams spread with data science, engineering, and operations capabilities. We also cover Petr Janda's recent podcast on the need for more engineering rigour for parity with other teams.