Anthropic Survey: 35% of Workers Expect AI to Handle Most Tasks Within a Year

·1 views

Anthropic's June 2026 Economic Index report, based on millions of Claude conversations and a survey of 9,700 users, found that over one-third of respondents expect AI to perform most or nearly all of their job tasks within 12 months. The study, titled 'Cadences,' is the fourth installment of the index Anthropic launched in early 2026 to track Claude usage across US states and hundreds of occupations. Usage data shows a clear workweek rhythm, with business-related conversations dominating weekdays and personal queries rising sharply on weekends. Higher-wage workers tend to use Claude more intensively outside standard hours, generating significantly more tokens and output per session than lower-wage counterparts. Anthropic notes the pattern suggests AI is augmenting human labor rather than replacing it, though younger workers surveyed expressed the greatest concern about job loss.

Read the full story at DEV Community

This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)

Developer Compares C# and Rust by Rebuilding a Real-Time Order Tracking Service

A software developer with extensive C# experience rebuilt a real-time order tracking service in Rust to compare the two languages hands-on, rather than relying on opinion pieces. The project used ASP.NET Core and SignalR for the C# version, and axum with Tokio for the Rust version, with both handling in-memory state and WebSocket updates. Informal benchmarks on the developer's laptop showed Rust's axum achieving roughly 34,000 requests per second versus 9,000 for ASP.NET Core, with idle memory usage of around 9 MB compared to 110 MB for .NET. The developer noted that .NET's JIT compiler narrowed the performance gap significantly under sustained load, making it competitive once warmed up. Key takeaways centered on Rust's compile-time memory safety, explicit concurrency model, and predictable tail latency, while C# was credited for its productivity and mature ecosystem.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Developer Spends 40 Hours Promoting App, Gains Only 68 Users in Three Weeks

An indie developer launched a convenience-store BOGO app and set a goal of 100 users without paid advertising. On launch day, a single round of posts across two communities yielded just 15 installs, many from personal contacts. Over three weeks, manual posting across five communities and multiple social channels produced small, short-lived spikes in downloads that never compounded. After 40 hours of distribution effort, the developer reached only 68 users and concluded that community posts cannot replicate the intent-driven timing of search traffic. The experience prompted a rewrite of the app's store listing using search-friendly language, and a broader reckoning with the limits of manual, unscalable promotion.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

How to Fix OpenTofu Registry Timeout Errors During Provider Initialization

OpenTofu users may encounter a 'Client.Timeout exceeded' error when running tofu init or tofu plan, caused by the tool failing to fetch providers from registry.opentofu.org within its default 60-second HTTP timeout. Common triggers include slow internet connections, firewall or proxy interference, DNS resolution delays, and occasional overload on the public registry. The most effective fix is setting the OPENTOFU_HTTP_TIMEOUT environment variable to a higher value, such as 120 or 300 seconds, a feature supported in OpenTofu v1.7.0 and later. Additional measures include enabling a shared plugin cache to avoid repeated downloads, verifying proxy settings, or switching to an alternative provider source if the registry is temporarily unavailable. For persistent issues in enterprise environments, configuring a local network mirror can eliminate dependency on the public registry entirely.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Go SDK Now Production-Ready for Building MCP Servers That Connect Claude to Your APIs

Developers can build Model Context Protocol (MCP) servers in Go to give AI tools like Claude Code structured access to internal APIs, databases, and CLI tools. The official Go SDK, hosted at github.com/modelcontextprotocol/go-sdk, reached production readiness in 2025, though its documentation still focuses on basic examples. An MCP server exposes tools, resources, and prompts to an AI client, with tools being the most commonly used feature in real-world deployments. Developers declare typed tool handlers with input schemas, and the model autonomously decides when to invoke them based on their descriptions. The SDK supports two transport options — stdio for local single-user tools and HTTP for shared or remote deployments — each suited to different authentication and scaling requirements.

0 comments Read more at DEV Community