Artificial Analysis tested 23 coding models on FrontierCode, which grades patches against maintainer criteria as well as tests.
Browsing Category
AI & Work
1522 posts
15 AI Meeting Transcription Tools That Can Make Notes Easier To Follow
A buyer’s guide compares 15 AI meeting transcription devices, led by Plaud Note Pro, and explains what to check on accuracy, access and storage.
Intel D270 GPU AI Accelerator–Up To 480 GB LPDDR5X
Intel describes the D270 as an air-cooled, PCIe 5.0 data-center GPU with up to 480GB LPDDR5X, aimed at long-context AI inference.
Dots, Bots, And Math: Rethinking AI, Jobs, And The Labor Zero Horizon
An essay links persistent agents, autonomous software and AI-assisted mathematics to a possible shift in how intellectual work is organized.
Recursive Self-Improvement Through Collective Intelligence
A headline points to recursive AI improvement through collective intelligence, but provides no article details to verify a specific development.
The Gap In Shared Understanding Of LLM Capability Is Widening
Andrej Karpathy estimates that access and professional use of frontier LLMs vary sharply, leaving most people with limited firsthand experience.
Impactful Scheduling For GPU Clusters
Ai2 says it replaced priority-based GPU scheduling with time budgets and fair-share allocation to direct scarce compute toward research priorities.
A relaunch with AI agents: how Gewerkton rebuilt its whole look in two days
Gewerkton replaced its whole look in two days, with AI coding agents carrying out the work from written plans: web app, phone app, Studio, icons, a website in forty languages and regional films. How it was checked, what broke and what the films cost.
How To Calculate The Cost Of Switching From Claude
Reports of Meta and Microsoft reducing internal Claude use highlight the costs companies should measure before switching AI models.
Microsoft Scales Down Its Internal Use Of Claude
A headline from The Information reports a one-third cut in Microsoft’s internal Claude spending, but timing, scope and rationale remain unverified.