We're a group of makers shipping together. We help each other stay accountable and reach our goals.
Apply for access
Marc Köhlbrugge
PRO
@marc
Anything that's configurable per-user complicates things. Examples:
Performance: Toggling that switch would need to wipe out the cache for all your todos.
UX: Requires some type of toggle that's hard to explain out of context (e.g. settings menu), but takes up too much space within the right context (next to the timestamp)
Code: yet another if/else statement (= two branches of logic)
I generally prefer to avoid per-user configuration.
I've explored many different ways to do implement streaks. The "as long as it's within 24 hours of the previous one" approach has its own issues. Just one example: if your last completed todo is at 3pm yesterday, and today you complete a todo at 4pm, you'd lose your streak. There's a bunch of other issues you'll run into when you actually try to implement it.
Fortunately, the current streak system works very well and has held up over time. It has support for time zones, etc. No need to change it I think.
The occasional confusion around time zones is that people don't always update them here on the website. In which case it's possible the todo is marked for a different calendar day than they'd expect. Seeing the time zone the todo was completed in helps clarify that.
( In case of your "Amsterdam-Tokyo" example you just make sure to complete a todo sometime on January 2nd (e.g. when you land). If you don't, you did in fact skip a calendar day so the streak being lost is correct behavior. )
I meant the 24h rule as an addition to the current rules (next day OR within 24h), but I havent tried to implement it so I believe you when you say there are more issues!
Would there be a meaningful difference between listing all your completed todos verbatim versus having AI rewrite them?
yes because you would ask the AI to filter out anything that--in its estimation--doesn't seem like it's on-topic / indie-hacking/shipping-related. And then also filter to the most-significant accomplishments.
I experimented with this a few months ago, but can't find any of the screenshots anymore 😅
Would you want a list of bullet points of what you completed that week or more of a regular paragraph?
Would you want it to be written from first perspective ("I worked on…") so you can share it with others? Or more of a summary to yourself by the AI? ("You worked on…")
@marc Since it's AI you could maybe let the user choose how they want it. Personally I think you should be doing a lot more to kick people in the ass to get shit done, the best coaching experience I've ever had was at a CrossFit box in Washington DC run by former Division-1 athletes who were all clearly on gear and all looked incredible, they really set the tone, the pace, the attitude to get everyone fired up and serious. I really think people would appreciate it, of course you have to be smart about it and not be a jerk or hurt people's feelings but if they can tell you're doing it to try to help them I think they'll appreciate it.
But to answer your question I would personally want it to score me, have it hook it up to productivity software like ActivTrak to get a sense of the amount of time I worked and GitHub to see how many lines of code I wrote. I want to know how I compare to the "pacersetters" who are actually successful, people like you and Pieter (and I think Ben is on his way). So maybe have it talking to me while impersonating a coach of my choice: Elon Musk, you, Pieter, etc.
@marc but yeah I think bullet points would be better than a paragraph
Would there be a meaningful difference between listing all your completed todos verbatim versus having AI rewrite them?
yes because you would ask the AI to filter out anything that--in its estimation--doesn't seem like it's on-topic / indie-hacking/shipping-related. And then also filter to the most-significant accomplishments.
You should probably remove the free trial. Especially if these trials costs you money to run.
The reason is that there's many people abusing free trials using virtual credit cards in order to try AI apps. I think there are communities where they share with tools which are vulnerable to this type of attack. You're probably on one of those lists.
Happened to me as well. I removed the free trial and costs plummeted and revenue (surprisingly?) shot up.
I didn't have a free trial earlier, so many people signed up but didn't pay. To improve the conversion I made the free trial, but I guess it's only those who don't want to pay, who start a free trial on the highest plan thinking they will create all videos before the trial ends. Each video creation costs me money, so I set the limit to 10 videos for free trials irrespective of the selected tier. Some see 10 videos limit and don't even create 10 videos maybe thinking their strategy failed.
I guess, there's no point in losing money. But I still don't want to feel regret about not letting users try before paying and hence losing out on customers. I guess I'll set the $1 trial, so the trial starts only if the card successfully makes a $1 payment. Will do this for a couple of weeks and see.
I checked people are using different cards from different banks, except for 3 accounts that use the same cc of Adam Bino which is not their name. Mostly these are debit cards, but there are also CC that fail. Stripe reports normal risk and no theft for these cards.
I guess for debit/prepaid cards I can turn off accepting payments in Stripe? For credit cards maybe I can start charging $1 at the start of the free trial to filter fake/stolen cards.
It's surprising to me that Stripe still has loopholes.
It will automatically hide 2 weeks after signing up