Connect with the Product Hunt community

It is so easy to fall into the trap of thinking that adding more text fixes bad outputs. We used to write massive instructions packed with edge cases and the model would end up ignoring half of them anyway.

Trimming everything down to clear bullet points and simple variables made the responses way more reliable. Cutting out the fluff really does force the model to focus on the actual task.

Do you prefer keeping your system instructions short or do you like adding strict guidelines for formatting?

That middle of the training data problem is so real. If you don't actively push for diversity every generated image ends up looking identical.

My main issue when adding random weather conditions or film stocks was that the product itself started changing colors or losing its original branding details.

Did you have to lock specific seed weights for the product while letting the background parameters randomize?

Searched the sub for agent-driven dev workflows and came up pretty empty. I mean those setups where autonomous agents write code, test it, and basically deploy small apps with minimal hand-holding.

I run a few small automated tools myself, so I'd love to swap notes with people who are actually building this way.

Is anyone else playing with this? What frameworks or orchestration setups are you running for your coding agents? More importantly, how do you keep them from losing context or breaking things when the codebase gets bigger? What decisions are you letting the agent make completely on its own without reviewing first?

Been on this platform a month now, and genuinely think the people here are sharper than what i've found anywhere else.
Curious if that's actually held true for others when it comes to hiring.
Have you hired someone you found on ph, or been hired yourself through here.
Also curious what your actual go-to platform is for finding people, ph, linkedin, twitter or somewhere else entirely.

Spent a full week forcing daily updates on social media because everyone says you have to do it for traction.

Ended up totally exhausted and only got 2 signups. Felt way better to just close the laptop and fix code in silence without posting about it.

How do you guys keep sharing updates without feeling like you're just performing for an empty room?

AI agents are getting better at finding products and comparing options but completing a purchase is a different challenge. Cart updates checkout forms shipping details and payment verification can all create friction along the way. Has anyone tested an agent against a real checkout flow? I'd love to know where it gets stuck and how much human intervention is still needed.

Question for folks documenting APIs, especially with docs in a separate repo from the code: how do you find out about changes before a customer does?

The big features usually reach us. What slips is side effects: a PR about a new filter that also lowers the max page size from 100 to 50, a default that flips, a new error code, a renamed webhook field. One line in a diff that's mostly about something else.

Our Net Promoter Score was consistently sitting at a depressing 12, yet our revenue retention was over 110%. We were completely confused.

When we jumped on 1-on-1 calls with customers, we discovered that users found our product's interface ugly and clunky, but the core data it generated was so indispensable to their daily operations that they couldn't switch to a competitor even if they wanted to. They were paying us every month while actively disliking the daily experience.

If you've ever inherited a product with high utility but terrible UX: do you pause all new feature development to rebuild the interface, or keep shipping value and live with a frustrated user base?

In the early days, every signup spike feels like validation. But the day your first paying customer cancels, your entire mindset shifts overnight. That single churn notification hurts more than a dozen missed signups combined.

It forces you to stop obsessing over top-of funnel acquisition and start looking at the uncomfortable truth product market fit.

To founders who have crossed the early stage hump: Does that sting of churn ever go away, or do you just learn to react faster to cancellation feedback?

We built a tool that people said looked great, felt smooth, and solved a neat problem. People signed up, paid for the tier, and then quietly drifted away. No angry emails, just silent cancellations.

That s when it hit us: we built a vitamin, not a painkiller. It was nice to have but when budgets got tight or schedules got busy, it was the first thing to get dropped from the daily routine.

If your product isn't tied to an urgent daily workflow, how do you pivot before churn drains your MRR?

I have so many ideas. So many. They arrive at random times, usually when I'm supposed to be doing something else, and they all sound incredible.

And then I look at my bank account and remember I have a budget. And the ideas are not free. They cost time, money, energy, all the things I don't have unlimited amounts of.

I'm preparing to launch my first product on Product Hunt next Monday, and I've been going down the rabbit hole of launch strategies lately

Some people say to launch right at 12:01 AM Pacific. Others say the day itself matters less than having a good product, a clear message, and people who genuinely care about what you're building.

As a solo maker, I'm trying to figure out where to focus my energy instead of overthinking every little detail.

For those who've launched before:

It's so easy to panic when a paying user asks for a specific workflow addition worrying they'll churn if you don't build it right away. But if you give in every time, your simple little tool slowly morphs into a heavy enterprise monster that nobody enjoys using. We started pushing back and asking what problem they were actually trying to solve instead of building what they asked for.

When a power user demands something complex, how do you politely say no without losing them?

It's so easy to get sidetracked by the loudest voices in your inbox asking for custom integrations or dark mode. For a long time we treated every request as a top priority just because a few users complained about it repeatedly. Then we realized people usually suggest fixes based on their own narrow workflows instead of looking at the product as a whole.

How do you decide which user feedback deserves a spot in your sprint and which ones are just noise you should ignore?

It s so tough to judge portfolios now when AI tools can make average work look polished. Traditional portfolio walkthroughs reward good presentation over real problem-solving.

We started doing a live audit of our actual app, asking candidates to fix real usability bottlenecks on the spot.

What design interview setup has actually worked for you lately?

I m building an app solo, and users can share their results as a link. The page has to open for people who aren t logged in, but the data belongs to a private account.

What I did: the link uses a random code instead of the real ID, and the page only gets the few fields meant to be public. Then I tested it in a private window, with a real code and a fake one.

It works, but I m not sure it s enough. If you ve built something like this, what do you wish you d checked earlier?

I started an apprenticeship recently, but the stack is mostly low-code platforms and builders where I don't even get to open VS Code. It s great for learning workflows, but I'm low-key terrified of forgetting how to write actual Python. I want to build a small AI or machine learning side project after hours just to keep my programming logic alive.

If anyone else has been in this spot what kind of small side projects actually kept you sharp without burning you out?

We spent three full months building a complex permissions dashboard because a few noisy users kept asking for it.

Turns out it just cluttered our settings and scared new signups away. We took a deep breath and deleted the whole thing back to simple roles.

Support tickets dropped instantly and signup usage shot up. Funniest part? the people who begged for it barely even used it.

How do you decide when a feature needs to be thrown away rather than fixed?

Strict time blocking is the fastest way to feel like a failure by Wednesday afternoon. Coding requires getting into a deep mental zone and getting pulled away for a 30-minute marketing task ruins the whole flow.

Dedicating a full day to just one major theme makes a huge difference. But my biggest struggle with day theming is dealing with urgent client requests that ignore my schedule.

Do you pause your theme day to fix fires, or push them to the next day?

Default image generation settings always default to the exact same boring composition and lighting.

When we were making marketing banners every batch looked like it came from the exact same template. Adding random focal lengths and lighting styles via script sounds like a smart fix to force some variety.

Have you tried playing with negative prompts as well or do you mostly focus on randomizing the main input parameters?

Saw that new app Tab raised at a huge valuation for a bot that lives entirely inside iMessage and WhatsApp.

It actually books stuff and spends money for you instead of just giving instructions.

Not having to download a new app is great, but texting feels super sketchy when real money is involved.

How can text-based tools give you clear receipts so you know they didn't mess up your order?

When evaluating new tools, I ve started ignoring flashy marketing pages and heading straight for the product's changelog.

Seeing steady weekly fixes and listening to real feedback tells me way more about the team's momentum than a row of testimonials.

Framing every update around a user headache like "fixed slowdowns on big exports" instead of vague tech speak makes a massive difference in showing care.

Has publishing a public release history helped your conversion rates, or do you find most visitors ignore it?

For the illustrative values 10, 11, 12, 13, 54, the mean is 20 and the median is 12. A single headline number hides that distinction.

Would you show a small distribution chart next to the summary, or expose the extra detail on demand? What design has worked for readers who are unfamiliar with statistics?

"Attention is loud and revenue is quiet" is such a spot-on way to put it.

We often fall in love with the launch hype on Product Hunt or Twitter, but that spike dies down fast. I had a tiny utility script I built in a weekend and completely forgot about.

Picking a flexible setup feels like overengineering when you're just starting out so we just threw everything into one flat database table to move fast.

It seemed like the right call at the time because we shipped our MVP in two weeks. But six months later when we needed to add relational data and user permissions, refactoring that massive table took an entire month of halted development. Sometimes moving too fast in week one creates a massive anchor in month six.

Did you stick with a simple database structure early on, or did you prep for scale from day one?