https://roboco.io/en/tags/gpt/ Recent content in Gpt on ROBOCO Hugo en-us Sun, 04 May 2025 11:59:09 -0700 https://roboco.io/en/posts/the-art-of-vibe-coding/ Sun, 04 May 2025 11:59:09 -0700 https://roboco.io/en/posts/the-art-of-vibe-coding/ <figure><img src="https://roboco.io/posts/images/Dohyun.png"><figcaption> <h4>Dohyun Jung - Principal Consultant, ROBOCO</h4> </figcaption> </figure> <p>Have you ever heard the term vibe coding? It didn&rsquo;t start out as a compliment. It was used with a hint of sarcasm to describe developers who wrote code relying only on intuition and feel, often ignoring rigorous structure or rules. Interestingly, though, it has recently come to be associated with generating code through prompts alone using artificial intelligence (AI) — large language models (LLMs) in particular. Of course, you can write perfectly exemplary, maintainable code even when using an LLM. And if you get code that works well on the first try, without a cycle of trial and error, that means vibe coding did its job — and that you genuinely understand what AI-driven development is about.</p> https://roboco.io/en/posts/gpt4.1-released/ Tue, 15 Apr 2025 07:22:08 +0900 https://roboco.io/en/posts/gpt4.1-released/ <figure><img src="https://roboco.io/posts/images/Dohyun.png"><figcaption> <h4>Dohyun Jung - Principal Consultant, ROBOCO</h4> </figcaption> </figure> <h2 id="tldr">TL;DR</h2> <ul> <li>The GPT-4.1 family is a developer-focused lineup built around coding, instruction following, and long-context handling.</li> <li>GPT-4.1 Mini and Nano are presented as solid options for work that needs fast responses and cost efficiency.</li> <li>From a vibe coding perspective, the key point is that Windsurf&rsquo;s free-usage event lets you try the models out without any financial commitment.</li> </ul> <h2 id="introduction---openai-releases-gpt-41-built-for-software-development">Introduction - OpenAI Releases GPT-4.1, Built for Software Development</h2> <p>In the early hours of April 15, 2025 (KST), OpenAI announced GPT-4.1, a new product family for developers. The family consists of GPT-4.1, GPT-4.1 Mini, and GPT-4.1 Nano, the smallest, fastest, and cheapest model of the three. These models improve on GPT-4.0 and can handle long contexts of up to 1 million tokens.</p>