Reflections on Over a Year of Using Large Language Models

askone.ai
2 min readOct 23, 2024

--

Hello everyone, I’m a software engineer with over a decade of experience. Recently, I launched my first AI product and would like to share my journey and thoughts with you.

When I first heard about ChatGPT from my friend, I didn’t pay much attention, thinking that AI was still far from my daily work. But one day, I had a chance to try ChatGPT myself and realized that AI technology had made tremendous progress. At that time, I was mainly coding in Golang while working on a small side project in my spare time. The frontend of the project was built with Vue.js v3, and since frontend development wasn’t my strong suit, the process was challenging.

It was during this period that Copilot became my lifesaver — I called it “The Synonym for Productivity”. With Copilot’s assistance, I significantly reduced the tedious work involved in frontend development, and the project progressed much faster. Although the project didn’t ultimately succeed, it was the first time I truly experienced the convenience AI could bring.

For my second project attempt, I relied heavily on ChatGPT to solve technical problems. I often had to tweak the way I phrased my questions and try different prompts to get the ideal response, which even led me to study Prompt Engineering. However, I found it frustrating to constantly COPY and PASTE while researching technical solutions. In this project, Copilot and ChatGPT were indispensable tools. Although the project failed again, I realized that AI had become deeply integrated into my work and life. Most of my time was spent on feasibility analysis, documentation review, and quick validation, and I desperately needed more convenient AI tools to improve efficiency.

Eventually, I embarked on a third project. Now, I’m proud to introduce askone.ai. The vision of this project is to create a browser extension powered by large language models, providing users with robust AI support while browsing the web. askone.ai supports multiple models, including GPT, Gemini, Claude, Llama, and Qwen. And offers a convenient way to interact with language models without disrupting the original design of web pages. It eliminates the need for frequent window switching and cumbersome COPY-PASTE operations.

askone.ai can extract keywords, generate summaries, translations, and suggest potential related questions based on the content that interests you. Additionally, users can engage in in-depth discussions with language models about segments of web content.

After 10 months of exploration and development, askone.ai was launched in September 2024 and officially opened for registration in October. I invite you to try it out and look forward to your feedback!

https://youtu.be/jqhx2ArBar4?si=5l_40jzZQ9qenD8k

--

--

No responses yet