
I really enjoyed working with Opus 4.6 in Copilot, especially for tasks that required a bit of...
I really enjoyed working with Opus 4.6 in Copilot, especially for tasks that required a bit of creativity (e.g. it was much better at building a landing page compared to GPT-5.3-Codex), but I never got a Claude subscription. I finally did, at last.
After spending 24 hours with Claude Code in VS Code, I've come to the conclusion that it is focused entirely on "vibe coding". It will absolutely do anything to get you to a working solution, even if it isn't best practice.
The code is pretty sloppy. It hardcodes things it should have created configs or constants for. When I asked it to update some sidebar design, along with a change that impacted a common element used globally, I expected it to respect my structure. I have things neatly organized in __sidebar.scss and __global.css, but instead of using each for its own purpose, it stuffed everything into the sidebar file.
It readily fills in the blanks for API calls. For example, while working with the Calendly API, it invented a couple of concepts that looked legit only until I tested them manually and they failed.
Also, the rate limits are insane; you can't just work off a Pro account. Compare this to the fact that I never hit my rate limits on Codex (not even close) or on Copilot, which I was using exclusively just a couple of months ago. People coding entirely on vibes would ask for even the tiniest of things using AI, so it makes sense to extract as much value from them.
Since the code is sloppy, I can imagine the state of the code in a few months. Even if you wanted to maintain it, you would need to ask AI to handle the smallest of tasks because it would be too tiring to comprehend such code.
I also understand now why people need so many prompt files (skills) and whatnot with Claude. It needs a lot of prodding and prompting and has difficulty working off clear instructions at once.
I will still continue giving a piece of my time to it for the next few days, but it seems to have not gotten off to a good start.
aiMost of us have seen a coding agent fail to complete a task we know it can do. We just don't...
googlecloudWhen building Generative AI applications, developers often encounter a massive bottleneck: sequential...
discussI’ve been thinking about sharing some electronic circuit posts on Dev.to — small circuits, DIY...
agentsWhat nobody tells you about exporting your multi-agent prototype to a local workspace. Every...
agenticarchitectAutonomous agents are genuinely good at answering messy business questions. Give one an LLM and a set...
aiPR volume went up, ticket quality didn't, and the gap got filled with LLMs on both sides of the review: bots reviewing, bots replying, bots occasionally arguing with bots about priorities that only existed in a teammate's head. Our CEO named the actual problem, and it's bigger than code review.
Workflows from the Neura Market marketplace related to this Stable Diffusion resource