Yazılar

Study Finds AI Tools Slow Down Experienced Software Developers in Familiar Codebases

/in /tarafından

A new study challenges the common assumption that artificial intelligence always speeds up software development. Conducted by AI research nonprofit METR, the study focused on seasoned developers working with Cursor, a popular AI coding assistant, within open-source projects they knew well. Contrary to their expectations, these experienced developers took 19% longer to complete tasks when using AI compared to working without it.

Before the study, developers predicted AI would speed up their work by about 20-24%, but the actual results showed the opposite. The study’s lead authors, Joel Becker and Nate Rush, expressed surprise at the findings, with Rush originally anticipating a potential twofold productivity increase.

These findings complicate the popular narrative that AI tools dramatically boost the productivity of highly skilled engineers—a claim that has helped fuel heavy investment in AI-powered software development products. While AI is often touted as a way to replace entry-level coding jobs, the METR study reveals that its benefits may not extend to all developers or coding scenarios.

Previous research has shown significant AI-driven productivity gains, with some studies citing up to 56% faster coding speeds or 26% more tasks completed in a given time. However, METR’s work highlights that these improvements might be more relevant to junior developers or those unfamiliar with complex codebases. Experienced developers, intimately aware of the nuances of mature open-source projects, tended to slow down because they spent extra time reviewing and fixing AI suggestions.

Becker noted that while AI-generated code was often on the right track, it frequently required careful correction to meet precise needs. The study authors emphasized that the slowdown was specific to the context of experienced developers working in familiar environments and might not occur in other development settings.

Despite the slower task completion times, most participants, including the study authors, continue to use Cursor, finding that AI makes coding less effortful and more enjoyable—comparable to editing an essay rather than starting from scratch. Becker explained, “Developers have goals other than completing the task as soon as possible. So they’re going with this less effortful route.”

Grammarly to Acquire Email Startup Superhuman in Strategic AI Expansion

/in /tarafından

Grammarly has announced an agreement to acquire Superhuman, an email efficiency startup, as part of its broader strategy to build an AI-powered productivity platform and diversify its business offerings, company executives told Reuters. Financial details of the deal were not disclosed.

Superhuman, known for its exclusive email tool and a lengthy waitlist for new users, was last valued at $825 million in 2021 and currently generates about $35 million in annual revenue. The San Francisco-based company is recognized for integrating AI features aimed at enhancing email productivity, with users reportedly sending and responding to 72% more emails per hour. The use of AI tools for composing emails on the platform has increased fivefold over the past year.

Grammarly, which recently secured $1 billion in funding from General Catalyst, has more than 40 million daily users and annual revenue exceeding $700 million. Founded in 2009, Grammarly is evolving beyond grammar correction and is considering a rebrand to reflect its expanded ambitions.

The acquisition of Superhuman follows Grammarly’s 2023 purchase of Coda, a startup that added AI-powered research, analysis, and collaboration tools to its suite. CEO Shishir Mehrotra described email as the next logical focus, noting that professionals spend roughly three hours a day in their inboxes, making email a critical communication and productivity tool.

Superhuman’s CEO Rahul Vohra will join Grammarly, along with over 100 Superhuman employees. Mehrotra emphasized that the Superhuman product, team, and brand will remain intact, continuing to serve tens of thousands of users. Vohra expressed optimism that the acquisition will provide Superhuman with greater resources to invest heavily in AI and expand into related areas such as calendars, tasks, and collaboration features.

Both leaders envision integrating Grammarly’s AI agents directly into Superhuman, creating a network of specialized AI tools that streamline workflows by pulling data from emails, documents, and other digital sources. This integration aims to reduce time spent searching for information or drafting responses.

Grammarly and Superhuman will join a competitive market for AI productivity tools, contending with established tech giants like Salesforce and numerous startups.

Anthropic Launches Claude Opus 4 AI Model Capable of Autonomous Multi-Hour Coding

/in /tarafından

AI startup Anthropic has unveiled Claude Opus 4, its most advanced artificial intelligence model to date, claiming the system can now code autonomously for hours — a significant leap in the evolution of long-context, reasoning-driven AI tools. The company also introduced Claude Sonnet 4, a smaller, cost-efficient sibling model designed for broader accessibility.

Backed by tech giants Alphabet (Google) and Amazon, Anthropic has carved a niche in building safe, high-performing AI assistants, with software development and autonomous task execution as core strengths.

What’s New with Claude Opus 4?

  • Autonomous task handling extended from minutes to multiple hours

    • Example: Opus 4 was used by Rakuten to code for nearly 7 hours continuously

    • Another experiment had it play a 24-hour session of Pokémon — up from just 45 minutes with Claude 3.7 Sonnet

  • Enhanced long-form coherence and persistent memory

  • Improved context retention, logic, and decision-making over extended periods

“For AI to truly have the economic and productivity impact that it can, models need to work autonomously and coherently for long periods,” said Mike Krieger, Anthropic’s Chief Product Officer.

Key Technical Upgrades

  • Models now toggle between fast responses and deep reasoning based on the complexity of the task

  • Integrated web search capability for real-time information retrieval

  • Claude Code, Anthropic’s developer tool for software engineering, is now generally available after a February preview

Strategic Context

The release comes in a week marked by major AI updates from Google and OpenAI, reflecting the intensifying race for AI supremacy. With Claude Opus 4, Anthropic positions itself as a strong contender in the high-performance, enterprise-ready AI space — particularly in software engineering, automation, and long-context tasks.

Market Implications

  • Strengthens Anthropic’s value proposition for enterprise use cases such as code generation, virtual R&D assistants, and simulation tools

  • Places pressure on rivals including OpenAI’s GPT-4, Google’s Gemini, and Mistral’s open-weight models

  • Reinforces investor confidence in Anthropic’s multibillion-dollar backers, as the startup moves toward fully autonomous AI agents