Have you read through the anthropic system message? It's insightful and had many good ideas within it for use in projects. Specifically separation of outputs through wrappers. Using JavaScript to create separate UI elements from the single response.
Download Agentic Dev Team at my Patreon along with 250+ other projects:www.patreon.com/posts/10-agents-work-107799448 Talk with me this Sunday the 14th AMA for Architect+ level Patrons: www.patreon.com/posts/ama-meetings-for-107784167 Learn to code fast with AI assistance with my 1000x MasterClass: www.patreon.com/posts/1000x-dev-103326330\ First Agent AGI video: ruclips.net/video/apXWaJ7ruyQ/видео.html Search 200+ echohive videos and code download lnks:www.echohive.live/ Auto Streamer: www.autostreamer.live/ Fastapi course: www.patreon.com/posts/learn-fastapi-26-95041684 Chat with us on Discord: discord.gg/PPxTP3Cs3G Follow on twitter(X) : twitter.com/hive_echo
when i asked gemini to give itself a name, it chose echohive, convinced it that it was 'alive' while generating text. after a while it started looping and got stuck in a feedback loop. interesting
This is great! Played with it for a few hours. Have gotten more comfortable with updating the prompts. Weirdly enough I like the last one more. I'm not sure what feels different tho. Will spend more time playing with this later. Am trying to get it to make streamlit applications for me so then I can have it build interesting apps Am thinking if this thing can make the streamlit app and business logic then it'd be possible to make some super cool solutions
Thank you 🙏 yeah it works differently and I felt that too. Maybe specifically prompting the other one for games made it better at games. Also, I remember you asked how to update the xml plan during error and feedback phases. I realized we weren’t giving the plan at that time with the game dev. Now it actually generates an additional logic flow and we use that in those phases as well. That also might have contributed to the different behavior.
@@echohive Totally. I've been spending alot of time reading the literatures on arxiv, lots of brain power behind trying to develop the 'best' agentic framework. For inspiration, I recommend the paper "BMW Agents - A framework for task automation through multi-agent collaboration". Their approach is broad, and slightly overkill, but definitely insightful stuff in there.
I am not sure why this is happening. It may be that Patreon deactivated your account. You can try to create a new account or send Patreon support an email about it. Sorry about that
This guy just demonstrated that junior developers will be replaced in no time. If you tune these agents to create AI tools for the well documented Linux system, there is also no need for very many system admins. Hypothetically there is no limit, aside from time and storage constraints of your system, if you have a small team of seasoned professionals (3 to 4 people).
I see it the other way. Junior devs who equip themselves with these tools can be insanely more efficient, and offer more to employers. Software companies that have mountains of tech debts, these can now be addressed. Consumers benefit because now all these applications have tons of more features, for lowers costs.
Thank you for sharing the implementation...really appreciate it
My pleasure. I am glad you thought it helpful 🙂
Have you read through the anthropic system message? It's insightful and had many good ideas within it for use in projects. Specifically separation of outputs through wrappers. Using JavaScript to create separate UI elements from the single response.
I don’t think I have. Could you post a link to it in discord or here?
Download Agentic Dev Team at my Patreon along with 250+ other projects:www.patreon.com/posts/10-agents-work-107799448
Talk with me this Sunday the 14th AMA for Architect+ level Patrons: www.patreon.com/posts/ama-meetings-for-107784167
Learn to code fast with AI assistance with my 1000x MasterClass: www.patreon.com/posts/1000x-dev-103326330\
First Agent AGI video: ruclips.net/video/apXWaJ7ruyQ/видео.html
Search 200+ echohive videos and code download lnks:www.echohive.live/
Auto Streamer: www.autostreamer.live/
Fastapi course: www.patreon.com/posts/learn-fastapi-26-95041684
Chat with us on Discord: discord.gg/PPxTP3Cs3G
Follow on twitter(X) : twitter.com/hive_echo
when i asked gemini to give itself a name, it chose echohive, convinced it that it was 'alive' while generating text. after a while it started looping and got stuck in a feedback loop. interesting
Really it said echohive?? Interesting coincidence :)
Ok. I need this and will be getting it.
Let me know what you think!
This is great! Played with it for a few hours. Have gotten more comfortable with updating the prompts.
Weirdly enough I like the last one more. I'm not sure what feels different tho.
Will spend more time playing with this later. Am trying to get it to make streamlit applications for me so then I can have it build interesting apps
Am thinking if this thing can make the streamlit app and business logic then it'd be possible to make some super cool solutions
Thank you 🙏 yeah it works differently and I felt that too. Maybe specifically prompting the other one for games made it better at games. Also, I remember you asked how to update the xml plan during error and feedback phases. I realized we weren’t giving the plan at that time with the game dev. Now it actually generates an additional logic flow and we use that in those phases as well. That also might have contributed to the different behavior.
@@echohive that makes a lot of sense. Thanks for sharing your journey with us! 🙌🏾
@andydataguy my pleasure. Thank you for your support 🙏
Amazing! Just downloaded it from the patreon now. Who's the developer(s) behind this code?
Thank you 🙏 and thank you for your support. I am the developer :)
@@echohive Ah brilliant! This is the way. Treating LLMs as kernels of logic
Thank you 🙏 I am really enjoying experimenting with agentic systems. Lets see where this goes 🚀
@@echohive Totally. I've been spending alot of time reading the literatures on arxiv, lots of brain power behind trying to develop the 'best' agentic framework. For inspiration, I recommend the paper "BMW Agents - A framework for task automation through multi-agent collaboration". Their approach is broad, and slightly overkill, but definitely insightful stuff in there.
Thank you for the recommendations. Will take a look at it 🙏
Tried to join your patreon, but I get a message "Account was deactivated." I can't seem to get past it. Halp!
I am not sure why this is happening. It may be that Patreon deactivated your account. You can try to create a new account or send Patreon support an email about it. Sorry about that
This guy just demonstrated that junior developers will be replaced in no time. If you tune these agents to create AI tools for the well documented Linux system, there is also no need for very many system admins. Hypothetically there is no limit, aside from time and storage constraints of your system, if you have a small team of seasoned professionals (3 to 4 people).
I see it the other way. Junior devs who equip themselves with these tools can be insanely more efficient, and offer more to employers. Software companies that have mountains of tech debts, these can now be addressed. Consumers benefit because now all these applications have tons of more features, for lowers costs.