Adding up the hidden costs of generative AI

For decades, developers have dreamed of a higher-level abstraction that would free them from boilerplate code. Generative AI comes closer than ever, with senior engineers already letting it take on more of the work. But the vision of an effortless AI-driven future is colliding with hard realities like climbing prices and alarming security issues. Could smaller, more specialized models be the way forward?

Top picks for generative AI readers on InfoWorld

Is AI the 4GL we’ve been waiting for?For more than 30 years, developers have awaited a true “fourth-generation language” at a higher-level of abstraction than Java or C++. With genAI tools producing code from natural language prompts, Matthew Tyson wonders if 4GL may have finally arrived.

Senior developers let AI do more of the codingThe stereotype says new coders are overly reliant on AI, but a recent survey revealed something different.

The era of cheap AI coding assistants may be overThe transition to an AI-driven utopia might be hitting a real-world constraint. With GPUs in short supply and major AI companies converging on similar pricing tiers, increased developer productivity comes at a price some cannot afford.

When it comes to AI, bigger isn’t always betterFor enterprises looking to unlock AI’s power while saving on costs, small language models, with a more specialized set of training data, may be the solution.

GenAI news bites

San Francisco AI technology conference draws protests

JFrog announces ‘agentic repo’ for AI-driven development

Microsoft signals shift from OpenAI with launch of first in-house AI models for Copilot

Swiss launch open source AI model as “ethical” alternative to big US LLMs

Google intros EmbeddingGemma for on-device AI

More good reads and generative AI updates

When AI nukes your database: The dark side of vibe codingVibe coders who let AI do most of the work are running up against a series of surprisingly familiar pitfalls.

The era of AI-generated ransomware has arrivedIt turns out cybercriminals aren’t so different from their corporate counterparts—and like most developers, they’re increasingly turning to AI.

‘Scamlexity’: We put agentic AI browsers to the test—they clicked, they paid, they failedAgentic AI-enabled browsers promise a magical user experience, but a digital security firm found the security risks are nothing new. Basic phishing scams top the list.