Tags: code generation

Lovable has released a mobile version of its AI-driven vibe coding platform on iOS and Android, letting users create web apps and sites by typing or speaking prompts. The app syncs projects with the desktop service, supports asynchronous builds, and lets developers inspect and tweak generated code. By keeping outputs web‑based, Lovable sidesteps recent App Store restrictions that have forced similar tools to redesign their delivery models. The launch aims to extend Lovable’s on‑the‑go workflow while preserving the core promise of turning natural language into functional software. Lire la suite

Anthropic has begun previewing a new "auto mode" inside Claude Code, offering a middle ground between the default safety‑first behavior and fully autonomous operation. The feature uses a classifier to allow Claude to perform actions it deems safe while steering away from potentially dangerous commands, such as mass file deletions or malicious code execution. Anthropic cites recent high‑profile AI‑related outages as motivation, and warns that the system is not flawless. The mode is initially available to team‑plan users, with broader Enterprise and API rollout planned in the coming days. Lire la suite

An AI model named Claude was used to create a new version of the open‑source library chardet. The process relied on metadata from earlier releases and on the model’s training on publicly available code, raising questions about whether the new code is a derivative work. Human reviewer Blanchard oversaw the output, but his involvement adds complexity to the legal analysis. The open‑source community is divided, with some citing the lack of a clean separation between the AI’s training data and the generated code, while others argue that a fresh rewrite constitutes a new work. Lire la suite

Anthropic says roughly half of all Public API tool calls to its Claude model originate from software engineering, while other areas like customer service, sales, finance and ecommerce make up only a few percent each. Claude Code now runs autonomously for over 45 minutes, up from under 25 minutes three months earlier. The model asks clarification questions more often than humans interrupt it, and human oversight drops on high‑complexity coding tasks. Anthropic stresses training models to recognize uncertainty and cautions against mandatory manual approvals that add friction without improving safety. Lire la suite

A developer turned to the Claude large‑language model to create a Python script that colorizes log output and supports scrolling in a terminal viewport. While the initial tool functioned, horizontal scrolling caused near‑full CPU usage on a single core. The developer asked the model for a zero‑CPU impact solution, only to learn that such performance is unattainable. Claude suggested low‑impact alternatives, but after extensive token consumption and code revisions, the effort stalled without a satisfactory fix. Lire la suite

A hands‑on comparison of Google’s Gemini 3 Pro and Gemini 2.5 Flash models shows that the higher‑tier model delivers deeper reasoning and smoother code generation for vibe‑coding projects, while the faster model requires more manual prompting and frequent fixes. The experiment highlights trade‑offs between speed and depth, with Gemini 3 Pro generally producing more complete results with fewer iterations. Lire la suite

A hands‑on comparison of three leading generative AI models—Google's Gemini 3 Pro, OpenAI's ChatGPT 5.1, and Anthropic's Claude Sonnet 4.5—was conducted by tasking each with building a web‑based Thumb Wars game. Gemini 3 Pro delivered the most complete and responsive code, automatically adding desktop keyboard controls and a 3‑D ring environment. ChatGPT 5.1 produced a functional prototype but required additional prompting for desktop support and lacked depth. Claude Sonnet 4.5 generated a playable demo with customization options but failed to implement promised keyboard controls. Overall, Gemini 3 Pro proved the fastest and most adaptable coder. Lire la suite

Inception, an AI startup focused on diffusion‑based models, announced a $50 million seed round led by Menlo Ventures with participation from Mayfield, Innovation Endeavors, Microsoft’s M12 fund, Snowflake Ventures, Databricks Investment, and Nvidia’s NVentures. Angel investors Andrew Ng and Andrej Karpathy also contributed. The company, led by Stanford professor Stefano Ermon, aims to apply diffusion techniques—traditionally used for image generation—to software development and natural‑language tasks. Its Mercury model has already been integrated into several development tools, and the team claims diffusion models can deliver higher token‑per‑second throughput and lower latency than conventional autoregressive systems. Lire la suite

Developers are increasingly turning to AI‑generated code, dubbed “vibe coding,” to accelerate software creation. While the approach mirrors the efficiency of open‑source reuse, experts warn it introduces opaque code, potential vulnerabilities, and weakened accountability. Security firms highlight that AI models often draw on outdated or insecure codebases, making it hard to trace origins or audit outputs. A recent survey found that a third of security leaders report over 60 % of their code now originates from AI, yet fewer than one‑fifth have approved tools for such development. The emerging risk landscape calls for new safeguards and clearer governance. Lire la suite

Berlin- and New York‑based startup Deta has introduced Surf, a beta AI‑driven application that blends web browsing with notebook‑style knowledge management. Users can browse URLs, summarize web pages, PDFs, and YouTube videos, and generate AI‑written code, charts, or mini‑apps within a Notion‑like document. The tool leverages multiple open tabs as context for its chatbot, supports offline work, and aims at students and researchers. Surf is offered for free with future plans for premium features such as cloud backup and multi‑device collaboration. Deta, founded in 2019, previously pivoted from an infinite‑canvas OS to focus on AI browsing. Lire la suite