ok artificial genius didn't know you enter user/pass in the backend
As mentioned before, BlackHatWorld is listed as "insecure" by Firefox and you are entering your user/pass. However with that being said, you bring up a fair point. If you want to enter your user/pass with SSL encryption, you can just go to https://af.articleforge.com/users/sign_in and you can log in using a completely encrypted environment.
Was glad to see the recent v1.5 update notification. Articles seem to have a tighter focus - or maybe I'm just getting better at choosing Primary and Secondary keywords.
I'd like to see Exact-Match enforced - or the ability to put keywords in quotes - to insure keyword phrase words aren't treated with ambiguity. For example, trying to generate an article on 'Article Forge' has led to a few stray sentences about metal-working forges. Or, an article about 'Article Generators' led to some random content about electric and diesel power generators...
I know you're on the BLEEDING EDGE of some very, very complex stuff. It's exciting to see the world of automated article generation move forward! A Multi-BILLION dollar market awaits! Some parts of the puzzle have largely been solved by universities and businesses: The basics of appropriate word substitution, spelling, grammar, and punctuation, etc.
I think your biggest challenge will ultimately be in STRUCTURING a document: To have truly cohesive paragraphs that begin with a topic sentence, followed by supporting sentences. Then, to create an overall article that starts with an Introduction/Overview, supporting Main Sections, and then a final Summary paragraph -- that kind of format.
I'm ALL-IN and going along for the ride, even if it's rough right now. Because I know these issues WILL eventually get solved.
I appreciate the comments! Nearly all of our focus right now is on exactly what you described, keeping things more tightly focused and improving the structure of content. A lot of that was realized in Version 1.5, but there is still a lot of work left for us for future versions!
In general with this new update it tries
really hard for exact match - with the case you mentioned (such as "article forge" or "article generators") a lot of that is because our program exhausts all of the information it can learn about "article forge" or "article generators" but realizes it doesn't have enough content to write a full length article, so it starts to research concepts it thinks might be related (such as "forge" or "generators"). We have spent a lot of time trying to code a separate research algorithm that figures out that "generator" might not be related to "article generator" and instead something like "content creation" might be more related (despite not sharing any words). We were hoping that more of this code was going to find its way into v1.5 but we found that for us to adequately add this support articles would start taking 10 minutes to generate as this "research" phase can be very long. So in Version 2 a big focus will be on making this research phase faster so that Article Forge can quickly figure out in entirety what a specific keyword phrase means and doesn't mean.
As for structure, I agree entirely with you. I think we can easily have AI that knows how to structure a document (and we have some working prototypes of that for other research projects), but the very hardest part for us is to take a fact that Article Forge researches or learns and figure out whether that fact is a "supporting fact" an "overview fact", etc. We added a
lot of code that tries to handle this logic in Version 1.5, which is where a lot of the increased "tightness" is coming from. However there is still a lot of room for improvement, and now that we have a framework that supports this type of thinking we are hoping to continue to address this in a very big way in Version 2.
Lastly we are working on some technology with WordAi (for Version 4) that can take a sentence and turn it into a "supporting fact" sentence or an "overview" sentence. This is great on WordAi's end as it allows us to completely rewrite sentences but we are very hopeful that this technology could later make its way to Article Forge, so that when Article Forge is writing sentences it can write them specifically to fit the type of discourse we need at that point in the article.
This tool would be perfect if you can add web 2.0 it would be a amazing!
If enough people keep asking we may add web 2.0 support, but for now we'd prefer to have Article Forge integrated into other tools (such as FCS Networker, GSA Ser, RankerX, etc.) that already handle web 2.0s very well. So right now if you have any of those tools you should be able to enter your Article Forge API info in those tools and it will automatically manage your Web 2.0s with Article Forge content.
