frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Westinghouse is claiming a nuclear deal would see $80B of new reactors

https://arstechnica.com/science/2025/10/westinghouse-is-claiming-a-nuclear-deal-would-see-80b-of-...
1•atilimcetin•1m ago•0 comments

AMZN Layoffs show what AI is coming for our jobs, whether its good enough or not

https://www.amazonclimatejustice.org/open-letter
1•todchavez•3m ago•1 comments

Trial investigates how smart flooring could protect from fall-related injuries

https://www.surrey.ac.uk/news/new-trial-investigates-how-smart-flooring-could-protect-older-adult...
1•1659447091•4m ago•1 comments

DataGrip Is Now Free for Non-Commercial Use

https://blog.jetbrains.com/datagrip/2025/10/01/datagrip-is-now-free-for-non-commercial-use/
1•achristmascarl•5m ago•0 comments

Daily-Verse

https://github.com/patlehmann1/daily-verse
1•lehmann_dev•9m ago•0 comments

Can a Startup Make Computer Chips Cheaper Than the Industry's Giants?

https://www.nytimes.com/2025/10/28/business/can-a-start-up-make-computer-chips-cheaper-than-the-i...
1•ripe•16m ago•0 comments

Applied Keynesian Psychology

https://fi-le.net/keynes/
1•fi-le•17m ago•0 comments

Granite 4.0 Nano: Just how small can you go?

https://huggingface.co/blog/ibm-granite/granite-4-nano
1•CharlesW•19m ago•0 comments

Volunteers Step in to Help Understaffed NOAA Track Hurricane Melissa

https://www.nytimes.com/2025/10/28/climate/noaa-volunteers-hurricane-melissa.html
5•perihelions•21m ago•0 comments

The spillover effects of AI on open source software development

2•jennjwang•22m ago•0 comments

Ask HN: AI has changed my job for the worse

5•yodsanklai•23m ago•2 comments

Bananas in smoothies block flavanol absorption by up to 84%

https://www.sciencedaily.com/releases/2025/10/251027224844.htm
1•Gaishan•24m ago•1 comments

Walking 15 min lowers cardiovascular risk by up to 2/3 compared to short walks

https://www.sciencedaily.com/releases/2025/10/251027224829.htm
8•Gaishan•28m ago•2 comments

Encryption using SSH Keys with age in Linux

https://ittavern.com/encryption-using-ssh-keys-with-age-in-linux/
1•birdculture•29m ago•0 comments

When O3 is 2x slower than O2

https://cat-solstice.github.io/test-pqueue/
1•keyle•30m ago•0 comments

An ex-Intel CEO's mission to build a Christian AI: Hasten the return of Christ

https://www.theguardian.com/technology/2025/oct/28/patrick-gelsinger-christian-ai-gloo-silicon-va...
6•teleforce•32m ago•1 comments

Simple Coding Tips: How to Make Your Code Last [video]

https://www.youtube.com/watch?v=anL8caCUWl0
1•josephleomoreno•34m ago•0 comments

SF Studios launches YouTube channel for Nordic film classics

https://www.broadbandtvnews.com/2025/10/17/sf-studios-launches-youtube-channel-for-nordic-film-cl...
2•gnabgib•36m ago•0 comments

AI psychosis is a growing danger. ChatGPT is moving in the wrong direction

https://www.theguardian.com/commentisfree/2025/oct/28/ai-psychosis-chatgpt-openai-sam-altman
4•devonnull•39m ago•0 comments

OpenFold3-Preview

https://github.com/aqlaboratory/openfold-3
1•arvindh-manian•40m ago•0 comments

Obstacles to the Complete Stance

https://meaningness.com/complete-stance-obstacles
2•santadays•43m ago•0 comments

API for LLM-enabled semantic data matching

1•Izaiaht1•48m ago•0 comments

Agent Security 101: stop sending your transactional data to telemetry services

https://www.ashpreetbedi.com/articles/agent-security
2•bediashpreet•56m ago•0 comments

Fil-C: A memory-safe C implementation

https://lwn.net/SubscriberLink/1042938/ac9a001023324bf8/
2•askl•56m ago•1 comments

Show HN: Noodle Seed – Create ChatGPT Apps for businesses to reach 800M users

1•uziiuzair•59m ago•0 comments

Israel strikes Gaza after accusing Hamas of ceasefire violations

https://www.bbc.com/news/articles/cgjdy5eevn2o
4•1659447091•1h ago•1 comments

From Keyboard to Voice: The Future of Computing Is Talking

https://zackproser.com/newsletter/2025-10-20
2•ingve•1h ago•0 comments

Show HN: Run PyTorch on CPU boxes, offload kernels to remote GPUs

1•medicis123•1h ago•0 comments

Preserving Digital Memory at the Festival of Floppies

https://digitalpreservation-blog.lib.cam.ac.uk/preserving-digital-memory-at-the-festival-of-flopp...
3•gnabgib•1h ago•0 comments

History of Lambda Syntax

http://www.hydromatic.net/draft-blog/2025/10/26/history-of-lambda-syntax.html
2•tanelpoder•1h ago•5 comments
Open in hackernews

My users say they'd cry if my app disappeared. None of them pay

13•Nek_12•8h ago
I built a habit tracking app for 3 years. I have users who say "I'd cry the entire evening" if it disappeared. One is Level 177 with 197 completions/week. Another hit Level 420 (42,000 XP or 10000+ habits completed).

None of them pay. Most won't respond to emails.

Here's the data and what I'm missing.

THE NUMBERS

Revenue: 35,238 installs over 3 years, $86 MRR from 20 subscribers, $4,091 lifetime revenue, 0.8% free-to-paid conversion, $905/mo burn (day job subsidized).

Retention: D1 4.2% (industry 40%+), D7 1.1% (industry 25%+), D30 0.9% (industry 15%+), 130% monthly churn.

The Paradox: 9 users reached Level 100+ (10,000+ XP). 6 still active, using it daily. 1 interviewed (PhD + PM): "I'd cry" if app disappeared. Payment status: 0 paying, 4 on expired promos, 1 can't pay (SWIFT blocked).

POWER USER ANALYSIS

Analyzed usage patterns for 3,000+ users by calculating coefficient of variation (CV = std dev / mean of daily completions).

Success archetype (User A):

- CV 0.33 (very consistent daily pattern) - 580 completions/month - 465 days active - Regular professional schedule - High cognitive load (PhD + PM) - Job-to-be-done: "Eliminate decision fatigue"

Failure archetype (User B):

- CV 0.86 (boom-bust cycles) - Still paying subscription - Stopped using 7 months ago - Quote: "Life is too irregular, alarms ring at wrong time"

WHAT I TRIED

Retention fixes (none moved D1 from 4.2%):

- Rebuilt onboarding 4 times (current: v4, takes users through creating first ritual + habit + completing it once) - Added AI coaching (GPT-5, personalized habit advice, agentic tool calls) - Added gamification (quests, XP, levels, streaks) - Built WearOS app (rituals on your wrist) (more for myself as it was half a day of work, but still) - Added analytics dashboard - Daily notifications - Early wins in first session (complete habit in onboarding), streaks

Features built that didn't help:

- AI chat (11.54% conversion when discovered, but only 2.4% of users find it despite being literally in the bottom navigation bar) - Journaling - Habit timers with on-screen overlays - Streak widget with goals and focus on small wins/future progress

Interview attempts:

- Emailed all 9 power users for 15-20 min interviews - Offered 6-12 months free Pro - Got 2 responses after multiple attempts - User C (Level 177, active daily): no response - User D (Level 148): no response - User E (Level 236, ADHD coach, agreed to call): ghosted - 3 others: no response after 2 weeks

Paying customer analysis:

- Queried DB: only 15% of paying customers match "power user archetype" (CV <0.5 or Level 100+) - 4 of 5 known power users DON'T pay, the other pays but doesn't use the app. - Pattern: random trial conversions, not established users - 30% of subscribers have 0 activity (forgot to cancel)

Cancellation reasons (analyzed all churned subscribers):

- 43% "don't use it enough" (avg 63 days before cancel) - 30% cost concerns (avg 33 days before cancel) - 12% found better app (avg 19 days before cancel)

WHAT I THINK I KNOW

1. Product works for people with regular schedules + high cognitive load 2. Fixed-time alarms break when life gets irregular 3. Power users love it but won't pay or engage 4. Random people pay briefly then churn 5. Feature additions don't fix retention (D1 stayed at 4.2% through AI, quests, WearOS, etc.) 6. No monetizable pattern after 3 years

THE QUESTION

Not asking what to do. Asking: What am I blind to?

I have extreme PMF for 9 people who won't monetize. Is this:

- A pricing problem? (tested $1-6/mo PPP) - A positioning problem? (marketed as general habit tracker, should niche down?) - A feature problem? (wrong features locked behind paywall?) - Unfixable? (category too hard to monetize - results take weeks, requires discipline users lack)

Happy to share queries, GA4 exports, whatever helps.

Comments

Nek_12•8h ago
Built this solo as a side project alongside day job. No external funding, 100% bootstrapped. $45k cash reserves gives me runway, but 3 years with these metrics has me questioning if there's a path forward.

Tech stack: Kotlin Multiplatform, Compose Multiplatform, targeting Android/iOS/WearOS. 99% crash-free rate, 4.2* rating. The app works - users just don't stick or pay.

The regularity pattern (CV) analysis was eye-opening: active power users have CV <0.5 (consistent daily routine), churned power users have CV >0.7 (boom-bust). But when I analyzed paying customers, only 15% match this pattern. Most revenue comes from people who try it briefly and churn within 1-2 billing cycles.

If anyone wants to dig into the data, I'm happy to share: - Anonymized GA4 exports - SQL queries for usage pattern analysis - Onboarding flow breakdowns - Retention cohort data

Posting this because after 3 years I'm stuck between "one more pivot" and "shut it down." Hoping someone sees what I'm missing.

leephillips•8h ago
I don’t grok all your lingo, so I may be missing something, but if you’re blind to anything it’s this: if users don’t have to pay, they won’t. No matter how valuable they say the service is to them.

At least one super-famous bestselling author tried to serialize his novel online with voluntary payments, and had to abandon the experiment. Nobody payed.

Nek_12•8h ago
Fair point. Two of my Level 100+ users have zero subscriptions - the free tier works for them. But here's the twist: I DO have paywalls (alarms, AI chat, analytics, infinite history). They just don't care about those features, or work around them. The one power user who IS paying (CV 0.86, irregular schedule) stopped using it 7 months ago but forgot to cancel. So the question becomes: are the wrong features locked? Or do power users just not need premium features?

I even ran an A/B test (hard vs soft paywall):

- Started by "recommendation" from popular entrepreneur - 3 months, 3k users - Baseline (soft paywall, can skip): 0.8% conversion - Hard paywall (must activate trial): 0.6% conversion - Revenue tracking broken (Firebase showed $0 both variants) - Killed experiment yesterday.

leephillips•8h ago
The only thing that occurs to me is the example of one of the few online services I pay for: Kagi search, for which I pay $10/month.

They have a bunch of features that I don’t care about and never use (AI, their web browser, ...). I just need their core search product (because Google is useless). Instead of putting features behind a paywall, they offer search free of charge up to a certain number of searches per month (I think). It’s enough to learn if their search is good enough to pay for. I pay to continue to use their core service more frequently (and I guess it’s the same with most of their paying customers).

Your customers seem to find value in your core service but don’t want to pay for the extras. Maybe you could apply the Kagi model?

Nek_12•8h ago
Power users like the Level 177 user are getting unlimited value from the core. They complete 800+ habits/month on the free tier. The extras aren't compelling enough to upgrade.

Kagi model applied to Respawn would be: free tier gets 50-100 completions/month. If you're crushing it like the Level 177 user (800/month), you pay to keep using the core.

This makes more sense than what I'm doing. I'm trying to monetize peripherals when the core is what they value.

The risk: does limiting completions kill the habit-building momentum? If someone hits their limit mid-month, breaks their streak, do they just churn instead of convert?

But worth testing. It's at least aligned with value - heavy users pay, light users stay free.

leephillips•8h ago
If you try this, please let us know how it turned out! Good luck.
jasonthorsness•7h ago
This attitude doesn't reflect poorly on the users. Outside any kind of charity situation, if you pay for something that is free it makes you a sucker. You need to make sure users understand that paying makes them savvy/smart or otherwise benefits them.
Nek_12•7h ago
Right now, paying for my app DOES make you a sucker - the Level 177 user proves it. She gets full value free.

The Kagi model Lee mentioned above would flip this: free tier is limited (50-100 completions/month), heavy users pay to continue. Then paying makes you smart (you're getting 800 completions/month for $5), not a sucker.

But that requires confidence the core is worth paying for. And based on my conversion data (0.8%), most people try it and decide it's not.

So maybe the real issue is: the core isn't valuable enough to limit. And if I can't confidently say 'this is worth paying for after 100 uses,' then I don't have a product.

jasonthorsness•7h ago
One thing I have observed (but this is in B2B) is that developers can be a lot more cautious around asking for money than they should be. Sales can often tell customers they need to pay and they will pay. Complaints about cost are part of the negotiation process in this environment. In B2C I don't know.
GamingAtWork•8h ago
I don't know, but could it be that your "pro" features aren't anything special? I mean, if a power user can be on the app for so long, and never actually need to use those features, why would they pay for it?
Nek_12•8h ago
This might be it. Free tier includes: habit tracking, ritual creation (no alarms), timers, journaling, basic history. Premium unlocks: alarms, AI chat, infinite history, analytics, extra quest.

Power users use alarms + timers heavily according to interviews. But two of them (Level 148 and 177) have zero subscriptions (never paid), so either they found workarounds or alarms aren't as critical as I thought.

The weird part: AI Chat has 11.54% conversion rate when users discover it (vs 0.8% overall). But only 2.4% of users even find it. So there's ONE feature that works, but it's buried and most power users never see it.

Maybe the answer is: make AI chat the core experience, not a locked bonus feature?

dukefound•8h ago
Congrats on building this and thanks for sharing. I'm working through a similar problem. One thing I'd want to know is how you're distributing the app - you're clearly able to get users. From your description the actual user archetypes aren't all that clear (apologies if you already mentioned). For example, what's the age band of user? What work do they do? Where are they located? What exactly are they using it for? If you're able to understand some of that then you may have a better sense of their behaviors and what features may resonate. In other words, I'm an advocate of niching down.
Nek_12•7h ago
Just pulled GA4 data. Demographics:

- Age: 40% are 18-24, 20% are 25-34 - Gender: 50/50 split (males have 2x longer sessions) - Location: Was mostly US/UK/Russia historically, but had an organic viral spike in Poland (June 25) that I can't trace - install referrer tracking broke. Source unknown (YouTube/TikTok?), ~2k installs. - Platform: 92% Android, 8% iOS (iOS converts at 6.88% vs Android 1.07%)

The 9 power users I described are from the pre-spike era. The actual bulk of my users now might be younger/different demographic from Poland, but I can't interview them to find out.

nicbou•7h ago
I experience the same with my website. Very few people donate, despite constant claims that everyone uses it all the time and that it made their lives so much better.

Such is life. I do the same with so many little tools that make my life better. Why pay for something if I can get it for free? There are so many things asking for my money. I say no whenever I can.

There are not many solutions here. Either sell your services, sell your users, or sell to your users. If people aren't upgrading to the paid version, the free version might be enough. You might need to adjust what they get for free.

Nek_12•7h ago
This resonates. I do the same - use free tools, only pay when forced.

You nailed it: 'If people aren't upgrading, the free version might be enough.' That's exactly my problem. Level 177 user completes 800+ habits/month on free tier.

So I either:

1. Lock core features - risk losing the 9 power users 2. Keep it generous - accept it won't monetize 3. Shut down

Question: on your website, did you experiment with limiting free usage? Or accept it's a free tool with optional donations?

nicbou•3h ago
My website is free by design; keeping it free has been my policy for years.[0] I ask for donations, but it was never a significant source of income. The website's birthday is a few days apart from mine, so I use the donations to get myself a really nice gift every year.

However, my business model is different. For software, you need to give people enough to fall in love with your product, but not enough to keep using it for years without paying.

[0] https://allaboutberlin.com/terms

saaaaaam•7h ago
Out of those 32k installs what’s your current number of daily active users?
Nek_12•6h ago
DAU is around 50 (GA4 shows '1-day active users: 50'). MAU is ~3k. So DAU/MAU is 1.7% - users open the app about once every 2 months on average.
Nek_12•6h ago
Actually, database says 10-20 (users who actually complete habits). And mau is 214, based on that metric then. but those are only signed-in users.
saaaaaam•6h ago
MAU isn’t a good indication for a habit tracking app which almost by definition requires regular engagement to deliver value. DAU is what matters.

How many of your DAU have used the app on 7 of the last 14 days, say?

Being brutal, if you’ve only got 10-20 daily active users after 32000+ installs then all the data analysis you’re doing is just hocus pocus.

You’re focusing on entirely the wrong problem. Your issue isn’t monetisation, it’s user acquisition and retention.

Nek_12•6h ago
You're right. 10-20 DAU from 35k installs is catastrophic.

Of those 10-20 DAU, probably 8-10 are the power users who use it 7/7 days (Level 100+ users, CV <0.5). So yes, sticky users exist but they're <0.1% of installs.

I'm not 'focusing on monetization' - I gave up on retention months ago. Tried:

- 4 onboarding rebuilds - AI coaching - Gamification (quests, XP) - WearOS app - Daily notifications - Early wins in first session

D1 retention stayed at 4.2% through all of it. Nothing moved the needle.

So I shifted to: 'Can I monetize the 10-20 who stuck?' Answer: No. They won't pay because free tier is complete.

You're saying I should fix retention first. Fair. But after months of trying and 0.0% improvement, at what point do you accept the category is broken or the product doesn't work for 99% of people?

The 'hocus pocus' analysis was trying to understand if the 1% who stick have patterns I can replicate. They do (regular schedule + high cognitive load). But I can't find more of them or get them to pay.

Tarsul•6h ago
(after rereading I don't have any new points but I will let my comment stay for posterity) If so many people leave your free app, consider why they leave instead of focusing on 20 people who stay (how much money could you even make with 20 people? Not enough.). Only then you can focus on monetization. Anyway, best of luck and grats for getting so many downloads (albeit not enough sticky ones).
saaaaaam•5h ago
I wasn’t trying to be mean when I said hocus pocus - what I meant is that the analysis isn’t really valuable because you’re focusing on such a tiny cohort that you could probably make it show anything.

If you truly believe that regular schedule and high cognitive load are the “sticky” that makes it work then you need to market to people using those points.

What’s your marketing strategy currently?

On the “when should I give up” point… who are your competitors?

Nek_12•5h ago
Fair point on the tiny cohort. You're right - 9 users isn't enough to validate an archetype.

Marketing strategy: None, really. 99% organic from Play/App Store discovery. Spent $1,300 on ads over 3 years, negligible results. No content marketing (tried, spent ~20k$ on an instagram account, gained 50 followers), no SEO, no outreach (apart from irregular reddit shills). Just ship features and hope App Store algorithms pick it up.

Had one viral spike in Poland (June 2025, ~1,600 users) but don't know the source - maybe a YouTuber review, maybe App Store featuring. GA4 misattributed it to paid ads which didn't exist.

Competitors: - Free: Habitica (gamified), Todoist, Google Tasks - Freemium: Productive, Habitify, Strides (3-5 free habits, pay for more) - Paid upfront: Streaks ($5 one-time) - Coaching: Fabulous ($70/year), Noom ($60/month) - Niche: Focus Bear (ADHD), Routinery ($4/mo, routines focused)

The irony: I think I accidentally built something that works for a specific archetype (regular schedules + decision fatigue), but I've been marketing it as a general habit tracker to everyone. Which explains the 4.2% D1 retention - 95% are wrong-fit users.

But I can't niche down based on 9 users who won't even respond to my emails.

saaaaaam•4h ago
Again, forgive me if this seems mean but are you writing your responses using AI? There can be many reasons for this, including English not being your first language, dyslexia, neurodiversity and many other things.

That aside (as it isn’t necessarily important) I think you need to find three things that represent the “sweet spot” for your app and try marketing around those.

Sadly I don’t think it’s possible to scale what is effectively a commodity app with plenty of competition without having a strong purpose, and without actively marketing.

I’m not sure todoist is your competitor. I use todoist. I use it because it allows me to create tasks and assign them to people and then it gets out of the way. I pay for the team plan (pro plan?) where I can assign to half a dozen folk.

Fabulous is very interesting - I got sucked in on a deal. Quickly realised it is VERY woo woo and aimed at a certain female focused audience. As a very non woo woo non-female I found the ‘coaching’ very not for me.