Super Synapse Update (Telemetry!)
Вставка
- Опубліковано 3 жов 2024
- 🚀🧠 Dive into the Future of AI with Super Synapse! 🧠🚀
Hey, Chatters! 🌟 Ever wondered what goes on behind the scenes of creating an AI that's not just smart but also understands you? 🤔💡 Join Joseph in our latest video as he unveils the exciting transformation from Professor Synapse to Super Synapse - making AI more personal, accessible, and fun for everyone! 🎉✨
🔍 What's Inside:
00:08 The big leap from Professor Synapse to Super Synapse 🚀
01:00 How Super Synapse becomes the experimental ground for innovation 🧪
02:00 Simplifying AI to make it your own - a journey back to basics 🛠️
03:00 Introducing telemetry: Your AI's thought process, now more transparent than ever! 📊
05:00 Emojis as a bridge to neuro-symbolic AI 🌉😄
07:20 Customizing your AI: From expertise to personality 🎨
[08:20 New commands for smoother interactions 📝
11:00 Super Synapse in action: Crafting a children's book together 📚🎈
14:20 Your feedback shapes the future of Super Synapse! 💬❤️
👀 Watch the full video to see Super Synapse in action and learn how you can make the most out of this innovative AI tool. Whether you're a tech geek, a creative soul, or just curious about AI, there's something in here for you! 🌈
👉 Don't forget to like, share, and subscribe for more updates on making AI your best buddy! 🌟💬
GPT: chat.openai.co...
Github: github.com/Pro...
www.synapticla...
@synapticlabs
We get a breakdown video on a tuesday 😮. What a great way to start the day. Thank you for sharing and open sourcing the Professor. It's been a game changer 🙌🏻. Looking forward to testing and working with Super Synapse. 💯
Thanks for this! Looking forward to testing!
Love using Synapse. It's really helping day to day with research.
Thanks for your generosity
Great job my friend!
It's been a lot of fun learning from your videos.
Professor Synapse is absolutely my favorite. I am testing Super Synapse right now and it does not give the telemetry code on top as Professor does. I love seeing that code after each prompt. How can I make it to give it to me in the beginning of each response instead of that link on the bottom? Thank you for your work.
Go into your settings/general and toggle on “Always show code when using data analyst”
Thanks for the amazing work! Professor Synapse typically ends the reply with a question to the user (typically with a couple of options on how to continue the conversation or how to assist further). I really like this aspect, since the options are usually good ideas on next steps. This seems similar to what ChatGPT on the mobile app does as well (more conversational).
However, Super Synapse does not seem to display this behavior and ends the answer in a very matter of fact way, not intending to continue a conversation. Is this on purpose?
PS - other feedback - Hidden Telemetry UX:
- I do prefer the hidden telemetry rather than seeing it every time (while knowing that if necessary I can peek under the hood). It makes the thread much cleaner.
- I guess that the "con" here is that, as a user, I don't "see" the "thinking" that Super Synapse is doing. As such, psychologically, it makes me feel like the reply takes longer than usual while I'm waiting for any words to show up - although for very good reason, since it's doing the telemetry. Just food for thought.
Great feedback! Not intentional on the lack of question…hmm lemme work on that I’ve been playing with some specific thoughts on this. I appreciate this immensely! And yeah I’m playing with the “hidden” telemetry this becomes less reliable in normal synapse since the prompt is so much longer and complex.
Smart guy
Thank You:):) This is Great!
I've shortened this to use a custom instructions, are you noticing any big performance benefit to use a custom gpt vs custom instructions?
Haven’t tried so let me know how it goes 😀
you did a great job in explaination. i would loved if you went more in depth or included more details.
However see you on May 7 yushiva🎉
What depth and details would you like? I am writing up some blogs on this too.
I LOVE Professor Synapse... but I'm left wondering: Why hasn't anybody developed a system/app that takes the API's from the top LLMs, created agents for each, and then have these agents all work together to brainstorm, debate, review, and solve problems? I often get 4 different answers from 4 LLMs, so why not have them all setup as agents "in one room" working together to come up with the "best" solution. I can't find anybody that's tried this... why not? Wouldn't having the "top minds" (LLMs) working together produce better results?
This exists already the problem is developing the architecture of “working” together and many companies don’t want to show how the sausage is made but you can check out tools like autogen, crewAI, mindstudio, voice flow and many others to build you own.
@@synapticlabs thank you for the reply! I'm happy to hear that it already exists, but I haven't been able to find it anywhere, scouring the internet, UA-cam, even asking all the llms if one exists and where to find it. Can you point me to any examples where someone is attempting to have LLMs interact with each other, for example Agent APIs from OpenAI, Claude, Perplexity, etc?
If you Google Microsoft Autogen or CrewAI you’ll find plenty
@@Nifty-Stuff apologies if I wasn’t clear. There is currently no reliable commercial off the shelf solution but you can build your own with the tools I recommended.
[ RE: Emoji usage ] As a product designer (ex-Instacart/Disney/Google) and private software developer, the introduction of symbols and pictographs (i.e., Unicode) in computer interfaces and language paved the way for your information architecture and design of Synapse. It is intuitive, inherent, and has precedent in interfaces going all the way back to the 1980s at Xerox which inspired the Macintosh interface (ref: Jef Raskin & co.).
When implemented correctly (i.e., sparingly and strategically)-as you have demonstrated in Synapse-Emoji can be instrumental in elevating the intuitive efficiency of a user interface and experience. Emojis are simply Unicode pictographs enhanced by color, depth, and contrast, which IMO makes them even more useful and effective when a user needs to decipher information through differentiation, hierarchy, and meaning.
Obviously, the presentation layer of an interface can be interpreted-and appreciated-in different ways, so Emoji usage is not one-size fits all as it boils down to a user's preference.
Hey man, I really appreciate the work that you did! I want to add custom instruction to my chat gpt. I am using free version. And mostly i use version 3,5 because its free. Could you send me a instruction so that i could write it in chat gpt. Thank you😊
everything is on my github but here's the custom instructions for super synapse. might not be so great on 3.5 though...
@@synapticlabs so what do you recommend for 3.5?
@@Phoenix-zm5gv unfortunately there’s not really a recommendation 3.5 is really really bad. You can use Claude instead but unfortunately the lower level models need 10x more prompting for 1/10th of the quality.
@@synapticlabs so you recommend using claude?
Can't wait to test this! Thanks for emphasizing the embodiment section of the prompt. I find Prof. Synapse helpful but some times the agent seems too playful in tone of voice. However, I want to be clear the thought process and assistance when working on an idea/challenge is very helpful. Great work!
Thanks Maxwell, that's exactly the point! The personality for Professor Synapse is 100% my preference, but I want folx to be able to take it and make it their own - excited to see what you come up with!