I like claude the best. I agree and that is what I have always said, these softwares unique value is how well they can utilize the LLM and make long helpful articles. This test is just to see what they produce with the same prompt, same topic, right out of the box. I agree they are only as good as the prompt given. Thanks for watching!
I think it is important to note that not really any real improvements were advertised except that it is going to be able to do GPT-4-level results much faster with slight improvements, but that it is able to do this faster handling to facilitate realtime interaction via the inbound voice and video interaction. Additionally it appears to be improved in language and mathematics skills, which seems to correlate with these superior spacial skills. It now essentially processes images and audio directly vs interpreting them as textual data, which brings this more close to the way humans process things - at some level. This is exciting in some regards, but I believe the Q* or GPT-4.5 or GPT-5 will be released somewhat imminently (within weeks) since they seem to have released this to deflate Google and to get Google and others to show their proverbial cards. After they are satisfied with seeing enough lain out by the competition I believe they're going to pop the new model on everyone; especially since they need to keep Plus et al users happy.
Watching some other SEO videos confirms your comment on the actual number of keywords "stuffed" in an article, it seems the tools that calculate keywords present in an article don't really tell the truth (they are more guessing, especially with all of the Google ranking changes).
You are requesting a markdown code, then surprised when it's putting the requested code into a code section, then can't find the copy code link/button on the top right, and posting a video about that 😐
you should include GPT 3.5 as well as zimmwriter. a neuronwriter score would not really be meaningful unless you ask the AI to include the keywords it uses to get higher scores.
I cant do zimm as I moved to mac. I am not interested in vpn to use it as I want even using it with a pc. I agree with not adding seo score. thanks for watching.
Is it possible to share the AI Content Grader GPT? Can't find it anywhere...
LLMs struggle to write if they don't have enough context or direction. I can drop a mega-prompt into Claude 3 and have it write over 2000 words.
I like claude the best. I agree and that is what I have always said, these softwares unique value is how well they can utilize the LLM and make long helpful articles. This test is just to see what they produce with the same prompt, same topic, right out of the box. I agree they are only as good as the prompt given. Thanks for watching!
I think it is important to note that not really any real improvements were advertised except that it is going to be able to do GPT-4-level results much faster with slight improvements, but that it is able to do this faster handling to facilitate realtime interaction via the inbound voice and video interaction. Additionally it appears to be improved in language and mathematics skills, which seems to correlate with these superior spacial skills. It now essentially processes images and audio directly vs interpreting them as textual data, which brings this more close to the way humans process things - at some level. This is exciting in some regards, but I believe the Q* or GPT-4.5 or GPT-5 will be released somewhat imminently (within weeks) since they seem to have released this to deflate Google and to get Google and others to show their proverbial cards. After they are satisfied with seeing enough lain out by the competition I believe they're going to pop the new model on everyone; especially since they need to keep Plus et al users happy.
Good insight. Thanks so much!
Watching some other SEO videos confirms your comment on the actual number of keywords "stuffed" in an article, it seems the tools that calculate keywords present in an article don't really tell the truth (they are more guessing, especially with all of the Google ranking changes).
You are requesting a markdown code, then surprised when it's putting the requested code into a code section, then can't find the copy code link/button on the top right, and posting a video about that 😐
No.... the markdown should convert to html and it didn't. Calm down... doesn't take away from the test.
I tried the new chat model in Article AI Generator, I was not impressed at all. Claude was better, The output in Claude is mush better
I wouldn't waste your time with Originality scores.
you should include GPT 3.5 as well as zimmwriter. a neuronwriter score would not really be meaningful unless you ask the AI to include the keywords it uses to get higher scores.
I cant do zimm as I moved to mac. I am not interested in vpn to use it as I want even using it with a pc. I agree with not adding seo score. thanks for watching.
@@LearnWire ah yeah, I forgot you downgraded to a mac
lol!