I've been using Claude Code, Codex CLI, and Gemini for software development. My feelings are:
* Codex has the highest quality generation
* Claude Code has the slickest and most feature rich CLI and
* Gemini lags on both, but is the best for architecture brainstorming.
@OpenAICodex utterly failed when trying to create a simple Svelte 5 app - even providing the Svelte 5 MCP and recommended https://t.co/ZLdbDhpses svelte text. I hope AI doesn't mean the end of meaningful advancement in frameworks.
Working with AI development tools. It's like being a kid at Xmas every day. And by that I mean the excitement but also anticipation and anxious waiting.
Very impressed by GPT codex terminal with the codex model. I've been using Claude Code for months. Claude has the better UI and tool mix, but GPT gets an edge due to a) its brevity - preserving context and b) right the first time code generation
Bun v1.1.43
- 1st-class S3 support
- HTML bundler
- V8 heap snapshots
- All but 1 of Node.js' "zlib" tests pass
- 100% of Node.js[ "path" module tests pass
- Many reliability & Node.js compatibility improvements
Thanks to 22 contributors!
https://t.co/SFilWl27kg
#openai o1 just solved a particularly tricky TypeScript Typing problem involving parameterizing types beautifully. I had spent the past hour trying to get Claude Sonnet 3.5 to come up with an answer without success. Chain of thought for the win.
I'm looking for some advice on a prompt to get either Anthropic Claude Sonnet or OpenAI O1 to optimize 500 - 600 lines of code without breaking the API. So far both models start hallucinating regardless of the prompts.
Less than 24 hours ago, Anthropic announced Claude 3. And some people are calling it better than GPT-4.
Some are "feeling" Claude 3 (Opus) has reached AGI level.
8 incredible examples:
I'm starting a new side project, https://t.co/ELrA7d0oYU, to connect technical podcasts with technical guests. Follow my progress as I #BuildInTheOpen at @CodersFM or at https://t.co/ZiSbHKgbv6