Four Thoughts About Using AI Effectively

Four thoughts about using AI effectively: AI models get better fast. Frontier AI models can solve today tasks that they were only mediocre at six months ago. It can makes perfect sense to use two different LLMs—for example, ChatGPT and Claude—on the same task, feeding one’s output into the other. For any serious work, you should get on a paid plan. Many people are not impressed with AI because they have only experienced freely available models. The difference between AI models on the free tier versus the frontier models provided by paid plans is huge. Anthropic’s Claude Code and and OpenAI’s Codex are not only for coding. Their ability to organize work, establish goal-oriented workflows, verify their work, and delegate work to subagents can be very effective for research, knowledge management, and other work. Find a tutorial that suits your case and become proficient using these tools. You can reply to this post by email.