Ai

Since we’ve established “AI Inspiration Collision Forum,” there’s been a lot of random content being created, with people experimenting with AI to record and publish things, but thoughtful reflection is becoming increasingly scarce. Moving forward, it would be beneficial to slightly control the output of this section and consolidate it into a monthly magazine format – releasing one article per month would suffice.

• Last month, we experimented with cursor, but due to the limitations of the free quota, we didn’t develop overly complex features; we just did some basic testing. We discovered then that Byte also released similar products, both using the same large models – Claude-3.5 – at their core. Byte’s product is called Trae, initially launched in the Mac version and finally released its Windows version in February of this year. Big companies are good because you can freely “white嫖” (literally translates to “free-eat”), without having to pay, with unlimited use of Claude-3.5 – this model performs quite well.

Artificial Intelligence (AI) has undoubtedly been one of the most hotly debated topics in the technology field in recent years, particularly over the past two years, with AI technology experiencing rapid advancements. Whether it’s deep learning, natural language processing, or computer vision and automated decision systems, the application scenarios for AI are constantly emerging. However, despite continuous technological breakthroughs, AI still faces a bottleneck similar to that of a Docker release – a lack of a killer app to truly ignite the market.

Ollama is an open-source AI tool designed to enable users to run and deploy large language models (LLMs) locally. Its goal is to provide a convenient and efficient way for developers to use models like GPT on their local machines without relying on cloud services. Ollama supports multiple models and focuses on optimizing performance, allowing even resource-constrained devices to smoothly run these models.

It seems like another year has passed, and the biggest change at work is a significant increase in AI participation. Previously, switching between different development languages required developers to be familiar with various language-specific API interfaces. Now, these basic code snippets can all be generated by AI, which is a huge blessing for developers.

ChatGPT

As early as 2023, I’ve written two simple introductory articles about it. Now it’s been 25 years – how to put this… I haven’t felt a significant improvement. It still needs to develop its own cognition, be able to reasonably break down tasks, and, of course, most importantly, identify whether AI-generated code contains bugs.

Two years ago, I added a copy function to the site, which took me half a day of tinkering. Ultimately, the rendering effect wasn’t quite satisfactory, and as a somewhat mediocre frontend developer, I didn’t plan on improving it – if it worked, it was good enough. This year, I used AI to develop a mini-program, which has also made me more familiar with frontend development. I refactored it (the AI redesigned it).

Just as we needed to learn the techniques of searching engines back then, we also need to learn some techniques for communicating with AI, providing reasonable and sufficient constraints, and efficiently obtaining the answers we need.

If you look at it from a different angle, current AI is like a very good student with excellent memory – it has the ability to memorize everything. What we need to do is learn how to communicate with AI correctly, effectively, and precisely describe our needs to help AI generate the expected results.

“GitHub Copilot” was released less than two years ago, and “ChatGPT” has appeared. I don’t fully understand the underlying principles; I’ve been using them for a while. The two tools’ levels of assistance are completely different, but they have both achieved a significant increase in productivity.