Combine the world's best AIs to answer your hardest questions.
Replies
Best
Hi Winson, it's a great tool for research but I need it differently for cms integration like wordpress and also there is almost 13-16% plagiarism in articles, please if you could add these I'll be highly appreciated.
@super_gaming_live_hd Ithy uses a lot of references, but every word it writes is unique! You can actually export articles as HTML, which should let you use them on Wordpress.
@super_gaming_live_hd If you have a link to the article, paste it here and we'll take a look :)
Report
The fin-tuning process refines the model's capability for specific tasks. Open AI customizes c GPTusing supervised training, where human ai trainers engage in conversions whiling bothe the user and the AI Assistant.
These trainers follow guidelines to ensure safe and meaningful interactions. The Model Learns from this interactive training to produce relevant and contextually accurate responses.
Report
Great job. I have used and recommended Ithy since it first started. I think Ithy deserves a big pat on the back.
@antonio_tena Thank you Antonio for the support! If you remember the first version of Ithy, it was very slow and had basically no features - thanks for sticking with us lol
Report
this just supper its giving all answers like a blog with everything its like i can see all the thing at once my mind thing of, there could be more in this but as of now this supper amzing
Fascinating approach. Technically speaking, how does Ithy manage context retention and coherence when aggregating outputs from different models with potentially varying architectures and token limits? Would love to understand the orchestration layer behind it.
Report
اینطورکه دقت کردم اکثر افرادیکه اینجا هستن تسلط کافی به زبان انگلیسی دارن، که در این دنیای مدرن مسئله بدیهی است، اما ممکنه در گوشهای از دنیا و در نقطهای فرد یا افرادی باشن که با اینکه شاید علاقه زیادی به همکاری یا همفکری داشته باشن اما بعلت عدم یادگیری زبان بین المللی از این خدمات محروم هستن و یا ممکنه بعلت خودکامگی رهبران سرزمینهایشان از این اطلاعات بهرهمند نیستن، بنظرم بهتره در این زمینهها فعالیت بیشتری انجام بشه، با تشکر از همه دوستان
Report
The hard part isn't combining outputs, it's knowing when to trust which model. Most ensemble approaches paper over that with a final summarization pass and call it done.
Report
Multi-model consensus is genuinely interesting for research-focused tasks. One thing I keep wondering about products in this space: when models disagree, how does Ithy surface that divergence rather than just averaging it away?
Report
The multi-model aggregation angle is interesting, but when models disagree, which one wins? That tie-breaker logic is probably where Ithy lives or dies.
Replies
Ithy
@super_gaming_live_hd Ithy uses a lot of references, but every word it writes is unique! You can actually export articles as HTML, which should let you use them on Wordpress.
Ithy
@super_gaming_live_hd If you have a link to the article, paste it here and we'll take a look :)
The fin-tuning process refines the model's capability for specific tasks. Open AI customizes c GPT using supervised training, where human ai trainers engage in conversions whiling bothe the user and the AI Assistant.
These trainers follow guidelines to ensure safe and meaningful interactions. The Model Learns from this interactive training to produce relevant and contextually accurate responses.
Great job. I have used and recommended Ithy since it first started. I think Ithy deserves a big pat on the back.
Ithy
@antonio_tena Thank you Antonio for the support! If you remember the first version of Ithy, it was very slow and had basically no features - thanks for sticking with us lol
this just supper its giving all answers like a blog with everything its like i can see all the thing at once my mind thing of, there could be more in this but as of now this supper amzing
Ithy
@abdul_rahim_rangrez_rahim_ Thank you Abdul! There's a lot of new features coming soon, so stay tuned!
اینطورکه دقت کردم اکثر افرادیکه اینجا هستن تسلط کافی به زبان انگلیسی دارن، که در این دنیای مدرن مسئله بدیهی است، اما ممکنه در گوشهای از دنیا و در نقطهای فرد یا افرادی باشن که با اینکه شاید علاقه زیادی به همکاری یا همفکری داشته باشن اما بعلت عدم یادگیری زبان بین المللی از این خدمات محروم هستن و یا ممکنه بعلت خودکامگی رهبران سرزمینهایشان از این اطلاعات بهرهمند نیستن، بنظرم بهتره در این زمینهها فعالیت بیشتری انجام بشه، با تشکر از همه دوستان
The hard part isn't combining outputs, it's knowing when to trust which model. Most ensemble approaches paper over that with a final summarization pass and call it done.
Multi-model consensus is genuinely interesting for research-focused tasks. One thing I keep wondering about products in this space: when models disagree, how does Ithy surface that divergence rather than just averaging it away?
The multi-model aggregation angle is interesting, but when models disagree, which one wins? That tie-breaker logic is probably where Ithy lives or dies.