SEOs keep sharing traffic screenshots for random sites from SEMrush and Ahrefs like it's accurate.
They’re not even close.
Here's an example?
This site had 63K users in GA last month.
SEMrush? ZERO.
In fact, this site died a while ago, according to the tool. 🙃
Oh come on, there is something charming about being clueless and staying up until 5am cold and alone while:
🙎🏻♂️ - "Still errors. Please fix it."
🙎🏻♂️ - "Still error. Fix. Think step by step."
🙎🏻♂️ - "no fix it again. We don't need 23 new node libraries installed either, please remove those"
🤖 - "Oh I see the issue now! 71 steps ago I deleted all of your .env data and removed all production ready code for test data. Would you like help with anything else? :)"
Oh come on, there is something charming about being clueless and staying up until 5am cold and alone while:
🙎🏻♂️ - "Still errors. Please fix it."
🙎🏻♂️ - "Still error. Fix. Think step by step."
🙎🏻♂️ - "no fix it again. We don't need 23 new node libraries installed either, please remove those"
🤖 - "Oh I see the issue now! 71 steps ago I deleted all of your .env data and removed all production ready code for test data. Would you like help with anything else? :)"
@BaniGrisson @PaulPromptChain @0xIlyy However, all the previous attempts to integrate it into Supabase remain in the code. As a result, in future tasks, lovable continues to work with the leftover garbage from Supabase, even though there was a request to remove it. It just snowballs from there.
@BaniGrisson @PaulPromptChain @0xIlyy For example, lovable creates an application and tries to use https-proxy-agent in Supabase, which isn’t possible. After several iterations, lovable builds a simple server and successfully integrates https-proxy-agent.
@billyjhowell@Replit Lovable easily gets stuck in loops, just like ChatGPT 3.5. After several rounds of "try-to-fix", you'll end up with code that might work but is cluttered with tons of unnecessary junk, which will interfere with future tasks :(
@BaniGrisson @PaulPromptChain @0xIlyy I gave a real case as an example. LLM have more knowledge, human have more intelligence for validate this knowledge. LLM is just a huge documentation database with an amazing search system.
@BaniGrisson @PaulPromptChain @0xIlyy yeah, sure, better and quicker than any human and then some shit tool like lovable or bolt try to run node-specific npm package on supabase edge functions xD
@marclou Thinking in a non-native language is impressive! I've tried many times, but even though I speak English, I still think in my native language =(
@eddiegonzales@sternenko То есть, то что ваши ебаные мопеды залетают на территорию белорашки, так что те вынуждены забивать их - то так, хуйня? Ох и долбоебушек, блядь. Чучело, лучше ищите своего пьяного Ивана, который опять левые координаты при запуске дрона ввел.