serverlessmania

Have you read the leaked source code of Claude Code? at this point they make sure it works with minimal bugs, don't think it is even tested by a real person, tests too should be vibe coded by agents!

I landed my first software engineering job from SO, and I didn't even apply to anything. I just had a profile with a decent reputation and never thought someone would check my profile and contact me for a role.

I miss the time when it really paid to be really good at a topic and help people. Today, nobody reaches out to you anymore, even in the workplace, everyone has Claude... feels depressing.

Thank you, StackOverflow. I'll be forever grateful!

And Alibaba is releasing the full model weights open source under Apache 2.0, Anthropic… fuck that company.

Why what's happening with Figma?

100% without forgetting the very "GOOD" idea of forking VScode, and thinking all devs will drop their tools and use the Fork :D

This product is going to be obsolete with the rise of terminal-based agents like Claude Code and Codex. Why would I abandon my entire workflow and toolset to use a fork of VSCode? Especially one that Microsoft will likely make increasingly difficult to maintain.

Developers use Vim, JetBrains, Emacs, VSCode, and many other tools—what makes you think they’ll switch to your fork?

Had an awful experience with Claude Code—it couldn’t even handle a basic API test for existing code, while Codex with o3 nailed it in one shot.

Claude Code tends to write meaningless tests just to get them to pass—like checking if 1 + 1 = 2—and somehow considers that a job well done.

GitHub acquired for 7b, Windsurf a VScode fork + Agentic LLM… 3b$. I should be missing something.

www.synthgenie.com – an AI-powered sound design assistant that connects to your hardware synth via USB MIDI and lets you control it using text, essentially turning the knobs for you.

You can’t honestly say you don’t fear death until you face it directly. Tell someone they’ll be executed tomorrow, then ask if they fear death, some people lose their minds when told they’re terminally ill. It’s not as simple as just saying, “I don’t fear death.”

“AI is misleading, there’s no actual intelligence.”

Oh wow, you figured out “Artificial Intelligence” isn’t literal. You should tell the rest of the planet. Maybe we should rename it “Statistical Pattern Prediction Machines That Are Better at Your Job Than You Are.” It’s a mouthful, but more honest.

“Modern tech hype mirrors medieval charlatans.”

Ah, yes, the classic “this thing I don’t like is medieval alchemy” comparison. Because if you can’t refute the tools, just compare it to snake oil. It’s cute how this guy thinks he’s the only one who’s noticed not every AI startup is curing cancer. The tech world has always had grifters. That doesn’t mean the tools themselves are fake—it means some people are dumb and others are greedy, which is kind of how humanity works.

“Hype is harmful and deceptive.”

Congratulations on identifying marketing exists. What’s next? You gonna blow the whistle on toothpaste commercials? Of course hype is driven by profit. That’s literally the point of business. But you know what else was hyped once? The internet. And electricity. And antibiotics. Being annoyed that people are excited about tech is not a worldview—it's a vibe. A really boring one.

"Less competent employees are often the ones promoting AI."

I don't understand how this post is HN front page…

"Another thing often missed by other frameworks are dedicated devtools"

From my experience, PydanticAI really nailed it with Logfire—debugging[0] agents was significantly easier and more effective compared to the other frameworks and libraries I tested.

[0] https://ai.pydantic.dev/logfire/#pydantic-logfire

Not all clients support that—currently limited to ChatGPT custom GPT actions. It’s not a standard. Fortunately, Anthropic, Google, and OpenAI have agreed to adopt MCP as a shared protocol to enable models to use tools. This protocol mainly exists to simplify things for those building LLM-powered clients like Claude, ChatGPT, Cursor, etc. If you want an LLM (through API calls) to interact with an your API, you can’t just hand it an API key and expect it to work—you need to build an Agent for that.

MCP allows LLM clients you don’t control—like Claude, ChatGPT, Cursor, or VSCode—to interact with your API. Without it, you’d need to build your own custom client using the LLM API, which is far more expensive than just using existing clients like ChatGPT or Claude with a $20 subscription and teaching them how to use your tools.

I built an MCP server that connects to my FM hardware synthesizer via USB and handles sound design for me: https://github.com/zerubeus/elektron-mcp.

Did something similar but controls a hardware synth, allowing me to do sound design without touching the physical knobs: https://github.com/zerubeus/elektron-mcp

Getting Amazon and k8s certifications will mean nothing with tools like this

Chrome extension to help me read and practice typing at the same time: https://github.com/zerubeus/type-selection

Pycharm AI agent (just started): https://github.com/zerubeus/aladin

Alexa skill to set Muslim prayer reminder, ask for prayer times, and so on: https://github.com/zerubeus/alexa_adhan

"Given a long enough time horizon, you'll deeply regret building on Serverless Functions"

Let's just apply "Hitchens's razor" to this kind of none sense:

What can be asserted without evidence can also be dismissed without evidence

NO!