Launched this week
ShouldBuild
Find out what your market wants before you waste months
19 followers
Find out what your market wants before you waste months
19 followers
You can build for months before finding out nobody wanted it. ShouldBuild reads what real users already complain about — across the app stores, Reddit, Hacker News and the open web — and hands you a build / don't-build verdict backed by quotes you can click. Keep the findings worth keeping, turn them into a concept and a backlog, and every Monday see what moved: new rivals, new complaints, and whether what you shipped made its complaint go down. 7-day free trial, no card needed.
This isn't bad. The tokens involved in doing the report is not going to amount to much. I'd be surprised if the tokens required per report amounted to $.50, maybe a dollar. So, out of the gate the pricing is on the high side for a customer but as a developer? Nice gross margins! :-)
The reports suffer, in my limited trial/opinion, from 'LLM helpful bias'. No matter how stupid the idea, ShouldBuild will try to find something. Consider this idea:
"a drone delivery service to service nuclear submarines with restocking uranium and pizza while in hostile waters. two for one on Wednesday"
Yes. I know, epically stupid on purpose to test helpful bias. The report back was don't build it and did flag illegal. It did, tho, try to send me down the path of heavy duty drone delivery with a bunch of next steps to chase down lawyers and users plus building a landing page to advertise the service. @itsarturodev , I'm sure you can see the results in your system along with the other two I ran.
To me, my single opinion, a quick 'illegal' check should stop the process in 30 seconds and say, this is illegal full stop, try again with something different, and be done. I think that adds to the creditability of the product's analysis vs a full report on praise and the like as ShouldBuild did here.
Again, my opinion is that if somebody is going to ask for objective, analysis before spending time and money; blunt, this is stupid/dumb/illegal is 100% valuable feedback. Yes, pivoting the idea to Drone delivery could be seen as trying to be helpful but the unseen risk is that the core product doesn't solve for LLMs helpful bias which may create issues for the so so ideas that get biased towards doing it when it shouldn't. I could be wrong.
Regardless, good luck with the product.
@rick_segal1 First of all, thank you for your valuable feedback, Rick.
I’ve read it all through in detail and have taken everything you’ve mentioned on board. It’s the most detailed and useful feedback I’ve received so far, so I really appreciate it.
As for the price, I believe it’s fair given everything the app offers and how similar tools are priced. It’s not just an analysis tool but also provides support and monitoring throughout the entire product development lifecycle. So having all the information consolidated in one place really helps with product development, refining the concept and being able to iterate on it.
Regarding LLM bias, what you’ve told me about being able to halt the pipeline when an idea is detected as unlawful is a great insight and contribution. I’ll take this on board and bear it in mind for future updates. Ultimately, the aim of an LLM will always be to assist humans, and the bias is imposed by its creators, but it’s true that I can adjust the system for such cases.
Apart from all that, have you had a chance to try out the other features? Do you have any feedback on them?
Thanks again for your message,
Best regards,
Arturo.
@itsarturodev
I'm looking through the rest of the product, yes. I have your email address and will send you some additional details which I hope can help as these PH threads are not the best place to talk shop!
And a second point: is social media still sincere in what it posts? Or is the only hope left in statistics?
@konstantin_tikhaev I understand your point, but I think we have to trust the reviews published online for any type of product or business.
Some will focus on inflating the statistics, but others will be 100% honest and have no malicious intent whatsoever. In any case, the LLM is designed to detect such cases, such as when someone posts a review ironically, intending to express the exact opposite. That is currently taken into account in ShouldBuild
@itsarturodev my question was broader, not just irony in product reviews. when public posting is performative, models just measure engagment tactics not "authentic" sentiment
The most interesting thing is modern comments of real users are mostly ai generated. So it's vicious circle of truth ) So, first layer should separate all comments and work only with "naturals"
@konstantin_tikhaev That’s a great idea, it’s true that with the rise of AI, there’s less and less UGC. I’ll bear that in mind and make a note of it to include in future updates; it shouldn’t be too difficult.
Thanks for your feedback Konstantin!
@itsarturodev My point was that behavoiral data is more reliable than text posts now because public commentary has become mostly "performance" rather than truth )
Always very valuable to know if an idea is worth building, as we can now build so fast.
@daniel_smidstrup Exactly, that's the idea! The barrier to entry is not code anymore, so there are plenty of ideas that can be executed, but not all of them are worth it!
Thanks for your comment, Daniel! :)
Beep! - New Era for Collaboration⚡️
Look like a cool product! I wish I had this when building my first products, would def help me make better decisions. Good luck Arturo!
@farid_sukurov Yes, I was inspired by all the people who are currently creating their products from scratch; I don’t want them to make the same mistakes that others of us may have made with our first projects.
Thank you for your comment Farid! :)