Stop searching for the best - just pick one!
At work, I had a browser tab group titled “AI” that had 6 tabs open; each one a different AI assistant. I was using the same prompt on each one of them trying to find the best AI who would get my focused attention and money - a paid subscription, once I committed to it. Eventually, I realised all I was doing was waste my time on something that barely mattered.
I hope this helps someone who is in a similar situation.
I realised that this is not as big of a deal as I thought it is . I’m not choosing my phone/laptop that i’ll keep for the next few years. There is no lock-in (unless you go for the paid annual plans) & I can switch at anytime. So I decided to pick one for now, get used to it & if I felt there is something better in the market after 2 months, I’ll happily switch to it without (hopefully) losing any context. I might lose the chat history, but I could still move everything else - my notes, learning, style etc. Instead, focus on the skill of using these tools well - prompting clearly, iterating thoroughly, asking the AI to ask clarifying questions & when to refine and trust the output.
The pace at which AI models update & releases a newer version is so quick. A thorough comparison done last month might not be relevant today. The model that was leading on the benchmarks might be completely crushed by a new player next month (yes, I’m looking at some of you Chinese models out there). This has happened multiple times already and the leaderboard keeps changing.
Treating these comparisons as durable fact is like navigating with last year’s map of a city that is actively under construction.
Compare once and make a decision. Stick with it or change it later, but don’t compare again without having used a candidate tool.
The parameters used for the comparison could also change. I looked at context window, price per input/output tokens, speed, features/capability set etc. But when the next update ships, these values could change making the comparison lose its merits. A model that was too expensive for your use case last month might now be quite reasonable today. Therefore, the important question to ask is, what do you need it for and what works well for that use case. Just pick one and go ahead with it. You can revisit your choice in 3 or 6 months, but not weekly/monthly.
The time & energy spent finding the “best” tool is almost always worth more than the cost of any single tool. The best tool is the one that you actually use to get things done & get you ahead!
If you are still wondering what those 6 tabs were, here: ChatGPT, Copilot, Claude, Gemini, Lumo & Confer.
I have not considered any open source models or tools that you can host yourself and run locally. While the principle remains the same, stop comparing and searching unless you are actively using any model/tool right now.