If you ask me who I talk to most frequently, you might be surprised by my answer. While I have a wife, kids, friends, and work colleagues, I find myself spending a considerable amount of time conversing with AI assistants. Claude is my primary go-to, followed by ChatGPT, but the field is rapidly evolving with promising newcomers like Mistral and Qwen making their mark.
My relationship with traditional search engines has shifted dramatically. Google has taken a backseat to Perplexity, which I now use for most of my online searches. The way I approach problem-solving has transformed entirely, especially in my professional work (I’m a software engineer/architect).
Take construction planning, for instance. I can show Claude my blueprints and get quick calculations for various scenarios:
However, I’ve learned to maintain a healthy skepticism. I never rely on a single AI’s response. Instead, I cross-reference answers across different platforms, pushing them to explain their reasoning in detail. The fascinating thing about AI is that the more you question, the deeper it goes into its reasoning. Some models, like O1s, can spend up to a minute processing a single question to provide more comprehensive answers.
While some AI services require payment, many powerful tools are freely available. For example, QwQ-32B-preview on Hugging Face offers extensive reasoning capabilities at no cost.
It’s crucial to understand AI’s limitations and tendencies. These systems often try to please users (perhaps to encourage subscriptions) by being overly positive about your work. I’ve learned to specifically request critical feedback, and surprisingly, it works! I’m particularly cautious about pricing information, especially for the Philippine market where online data is often scarce. In these cases, I either provide my own pricing data or use Perplexity to search for current market rates.
Through regular interaction with AI, I’ve developed an intuition for what questions will yield reliable answers and which might lead to hallucinations. I’ve learned to identify the most effective inputs and the tasks where AI truly shines. At this point, I can’t imagine working without it – it’s become my enhanced cognitive companion.
However, I want to emphasize that my goal isn’t to replace human expertise. Instead, I use AI to deepen my understanding and formulate more informed questions for professionals. While AI significantly accelerates the learning process, it shouldn’t be your sole source of information. It’s a powerful tool for augmenting human knowledge and decision-making, not replacing it (yet).
I’ve found this hybrid approach – combining AI assistance with human expertise – to be incredibly effective for learning and investigation. It helps me learn concepts more quickly and engage better with professionals in their respective fields.
Quick Links
Legal Stuff