A new AI journey! (content automation)

[Waifu_tatyana]

These inspiring AI journeys makes me wish I learned some coding and NLP seriously. It's probably too late to start coding from scratch lol.
Good luck on your journey mate

 

[HigurashiCoder]

Good luck with your journey.

 

[Talyn]

These inspiring AI journeys makes me wish I learned some coding and NLP seriously. It's probably too late to start coding from scratch lol.
Good luck on your journey mate

You can definitely learn enough python to do great things with openai, even do the same things as I try to implement, if I can do it, you can do it!

 

[Waifu_tatyana]

You can definitely learn enough python to do great things with openai, even do the same things as I try to implement, if I can do it, you can do it!

Yeah I'm trying to learn bits about openAI api and Rest api. If you don't mind can you point me to some good resources I can start with. I have minimal knowledge in coding (I know C and data structures to some extent).

 

[BroTherm]

The best way to learn fast if you already know some coding is to start doing stuff. Set up a Wordpress site, install python on your PC and start writing scripts.
Creating with the openai API and posting to Wordpress an article is like 50 rows of code.
Google is your friend in each step you don't know.
Also browse around and you will find half of the stuff already written and shared.

 

[Jammy0308]

- the keyword cleaner, a brute approach with spaCy, even with pipelines, won't work or would need serious power when you reach keywords lists with a few k inside, as I was working on pairs, you can easily go into 100-150m similarity checks, which is insane.

Why not use Openai embeddings to test for similarity? no need to run locally then.

 

[Talyn]

Why not use Openai embeddings to test for similarity? no need to run locally then.

I got cocky... I guess, but indeed to get the embeddings from ADA-002, then calculate the cosine and get rid of everything with a similarity >0.90-0.95 would be way faster, I'll give it a try

could also use some clustering as well with sklearn, but I feel I'm doing way too much things at once

 

[Talyn]

Interesting, my shitty second website, without a proper Re3 is getting more traction than my last website with a real good Re3 implemented that pass AI detectors.

One website is in a odd niche with 3k articles, 2nd one is in a women targeted specialized niche with 850 articles.
AI website isn't magic, publishing volume won't make anything happen if I don't pick the right niche.

 

[Razorbl2]

Interesting, my shitty second website, without a proper Re3 is getting more traction than my last website with a real good Re3 implemented that pass AI detectors.

One website is in a odd niche with 3k articles, 2nd one is in a women targeted specialized niche with 850 articles.
AI website isn't magic, publishing volume won't make anything happen if I don't pick the right niche.

What traffic are you getting?

 

[MoonStarz]

You being doing some serious work, sometimes I can barely get a project done because I got distracted easily or I'm to lazy to work, good to know you are doing better and improving your skills wish you the best of the luck and keep it up.

 

[Acquire Backlinks]

Good luck with your journey... What will be your budget for this project?

 

[sanators]

Good luck on the way up.

One website is in a odd niche with 3k articles, 2nd one is in a women targeted specialized niche with 850 articles.

are all these articles written with the help of chatgpt? if yes, you are moving fast..."writing" a lot.
also did you buy aged domains? or new ones?

 

[Cripto King]

Awesome thought good luck

 

[jamesj65]

Good luck with your journey, and best wishes!

 

[Talyn]

What traffic are you getting?

Not enough, 10UV/day - My GSC CTR is between 2-3%
So I'm reworking the algo content generation, when I read the articles, they seem quite good, but maybe I'm doing something wrong? or I should improve the design?
I have a website that is under the generation process, only 14 articles so far, but this time my keywords/article list is way better, instead of bruteforce it -> PAA generator -> titles -> outline -> Re3, I implemented keyword analysis Volume/KD/KGR sorted it by difficulty and improved the title/outline generation, we'll see in a month.

That's what annoys me, you can't make conclusions right away, you need to wait a few weeks to see if it gets any traffic/traction.

You being doing some serious work, sometimes I can barely get a project done because I got distracted easily or I'm to lazy to work, good to know you are doing better and improving your skills wish you the best of the luck and keep it up.

I can't find the youtube video that summarize it, but if you don't think you CAN do it, you WON'T do it, because if you think you can, you'll find ways to learn/do it/stay motivated.

Good luck with your journey... What will be your budget for this project?

So far I'm under$300 with hosting, domains, tokens etc.

Good luck on the way up.

are all these articles written with the help of chatgpt? if yes, you are moving fast..."writing" a lot.
also did you buy aged domains? or new ones?

100% openai GPT-3.5 turbo generated
I have a mix of brand new/expired. So far I only used brand new, but the website I'm generating now is using a recently expired clean domain


--------------
Small update: I'm currently working on a pin/instagram picture generator and a bot posting signals.

 

[Talyn]

Small update, So far:
- $240 spent in openAI
- $100 in domains
- $140 for hosting

Not really picking up in term of traffic. I will focus on social signals and backlinks to see if it improves or if I'm doing everything wrong.

 

[Fardin2012]

Hi Mate, I also started something like this but I already generated huge amount of articles, basically I scraped keywords from semrush but not cleaned it, that why I see many same meaning articles, do you think is this a big issue? any advice on keywords cleaning?

 

[Talyn]

Hi Mate, I also started something like this but I already generated huge amount of articles, basically I scraped keywords from semrush but not cleaned it, that why I see many same meaning articles, do you think is this a big issue? any advice on keywords cleaning?

You can use ada-embeddings model, then calculate the cosine to get rid of everything that has a similarity > 0.90

Then you can even cluster your keywords with sklearn.

Anyway some people don't care about duplicate content and get way better results than me.

 

[Talyn]

Websites 4, 5 and 6 are up with respectively 884, 1456 and 1564 articles that are 1600-3500 words long with pictures and videos.

I spent $100 on backlinks, to see if it influences traffic.

 

[Zatoichi]

@Talyn , what is your process for pictures? Where do you get the pictures from and how do you insert them into articles? I find it very difficult to speed up pictures process as I have to do it manually. Main concern is for the picture to be relevant to the article, and preferably to the subheading of the article where it is placed.