Best AI Tools: An Opinionated Guide to What to Use

The shift from general chat to specialized tools marks a quiet but profound evolution in how people use artificial intelligence. A year ago, many guides to AI use focused almost exclusively on large language models (LLMs) like ChatGPT, Claude, and Gemini, primarily for text-based interactions. Now, a leading voice in AI education points to a landscape dominated by purpose-built applications for specific tasks, signaling a maturity in the AI ecosystem beyond simple conversational agents.

Ethan Mollick, a professor at the Wharton School known for his "One Useful Thing" newsletter, recently updated his popular guide to AI tools, moving away from a broad endorsement of general chat models. His latest recommendations now highlight tools like Gamma for presentations, Midjourney for image generation, and Perplexity for search, among others. This update reflects a broader trend: users are increasingly seeking out AI that excels at a single function rather than trying to force general-purpose chatbots into every role.

From Chatbots to Specialists

This change matters because it redefines what "useful AI" means for many people. Previously, the advice often centered on mastering prompt engineering for a single chatbot, aiming to coax it into performing diverse tasks from writing emails to brainstorming ideas. The new emphasis suggests that better, more efficient results often come from using an AI specifically designed for a particular job, such as generating code or summarizing documents, rather than relying on a general model to handle everything. This specialization mirrors the development of software itself, moving from monolithic suites to a collection of best-in-class applications.

For everyday users, this means less time struggling to make a chatbot perform a task it wasn't optimized for. Instead of trying to get ChatGPT to create a polished presentation with embedded visuals, users can now turn to tools like Gamma, which are built from the ground up to handle presentation design. Businesses can leverage these specialized tools for improved efficiency in specific departments, from marketing content creation with image generators to research with AI-powered search engines, potentially leading to higher quality outputs and reduced manual effort in targeted areas.

The AI Toolkit Expands

This development fits into the larger narrative of the AI race by demonstrating a diversification of offerings beyond the foundational LLMs. While companies like OpenAI, Google, and Anthropic continue to push the boundaries of general AI capabilities, a parallel ecosystem of specialized AI applications is thriving. These niche tools often integrate with or build upon the underlying power of major LLMs but package that power into more user-friendly, task-specific interfaces. It shows that innovation isn't just about bigger models, but also about smarter application.

One concrete thing to watch in the coming months is how quickly these specialized AI tools begin to integrate more seamlessly with each other. The next phase might not just be about having a tool for every task, but about having a suite of specialized AIs that can effortlessly share data and workflows, creating a truly intelligent, interconnected toolkit for work and creativity. The future of AI utility may lie less in a single all-knowing agent and more in a finely tuned orchestra of digital specialists.