General
p/generalShare and discuss tech, products, business, startups, or product recommendations
trending
•

3mo ago

Is the Claude connector directory a real acquisition channel, or just launch-week spikes?

Keep seeing the same playbook this week: expose your product as an MCP server, add OAuth so it connects in one click, submit to the connector directory. The claimed numbers are always impressive. The latest one: 109 visitors, 41 signups, 13 paid trials in a weekend, $0 spent.

I haven't tried it, so I can't tell how much is real and how much is the usual launch-week spike that dies by Wednesday.

It doesn't feel crazy though. More of my clients' traffic is coming through AI tools every month, so distribution inside them has to start working at some point.

Anyone here actually listed their product as a connector? Did the users stick?

•

3mo ago

Why couldn’t I find a single decent dev tool to help me vibe-code a website?

I ve tried more than 10 AI website builders from well-known platforms to startups, and from free tools to paid services but I always end up battling endless bugs. They seem great at building an MVP, but not a genuinely polished website. Sometimes, the AI agent doesn t even understand what I m asking it to do. People might say my prompting skills are the problem, but when I give the exact same instructions to ChatGPT or Claude, they understand them perfectly. Am I missing something? Has anyone found a tool that actually works well for vibe coding a polished website? Someone, please help...

•

3mo ago

Building got easier after I launched

Before launching, almost every problem felt important. After putting the product in front of real people, it became much easier to see what actually mattered and what I had just been overthinking.

Still very early, but I m starting to feel like launching is less about saying the product is finished and more about finally getting better information.

Did your priorities change after your first users came in?

•

3mo ago

A voice model can score well and still feel terrible to use. Why?

I work on both ASR and TTS evaluation, and one thing I ve learned the hard way is that model quality and product quality are not the same thing.

You can have good recognition accuracy and still get an output that makes the user think: I would ve been faster typing this.

The failures I notice most aren t always obvious transcription errors.

Sometimes every word is technically correct, but the system misses the intent.

•

3mo ago

For those testing Siri AI on iOS 27: what works, and what’s still missing?

We re a couple of months into the iOS 27 beta now, and the broader release is getting pretty close.

For anyone who s been using the new Siri with Apple Intelligence: how has it been?

•

3mo ago

Anyone else asking Claude before Google now, but only for the messy stuff?

Noticed a weird split in my own habits lately. For anything I can phrase as a clean question, Google (or ChatGPT tbh) still wins, it's faster and I trust the top result.

But for messier stuff, like "I have half a workflow in n8n, half in Zapier, and I want to merge them without breaking my CRM sync", I go straight to Claude now. Not because it knows some secret answer Google doesn't have, but because I can actually paste my whole mess in and think out loud with it instead of translating my problem into a search query first.

•

3mo ago

What's your preferred visual style for an AI agent SaaS product?

I'm currently looking to revamp the product's visuals. The previous style featured robots exploring planets, using a lot of purple, which was a bit too cutesy. I've seen many websites with a minimalist style, but just black and white lines seems too monotonous. What design styles do you think are trending this year? If you were the design lead for an AI agent, how would you design the website?

•

3mo ago

Where should autonomous AI reasoning stop and human authorization begin?

I ve been working on a problem that keeps becoming more important as AI agents gain access to tools and real execution capabilities:

Capability and authority are not the same thing.

An agent may be technically capable of modifying a record, triggering a workflow, calling an API, or taking another consequential action.

But should the ability to perform that action automatically imply that the agent is authorized to do it?

Is "personalization" in marketing just surveillance with better UX?

Every new AI marketing tool leads with "hyper-personalization" dynamic content, predictive recommendations, data-driven targeting. But I keep seeing the same disconnect: tools pitch this as better customer experience, while actual customers describe eerily accurate targeting as unsettling, even when it's useful.

Feels like the industry optimizes for "relevance" because it's easy to measure, while "trust", which might move the opposite direction and gets ignored.

Curious how people building or using this stuff daily see it:

Marketing a screen-time / focus tool on platforms built to trap your attention feels like a paradox

Trying to reach people who want to reduce their screen time or build better focus habits is honestly a weird loop. Every place where indie makers hang out (Twitter, Reddit, PH) is engineered for maximizing engagement. For those of you building tools around digital wellbeing or focus: how are you actually finding your first users without contributing to the exact noise you're trying to help people escape?

First
Previous
•••
498499500
•••
Next