USEFUL Agentic Workflow: AUTO-Updating Blog with Claude 3.5 Sonnet
ฝัง
- เผยแพร่เมื่อ 7 ส.ค. 2024
- What if you could build an Agentic Workflows that worked while you slept?
Discover how to create powerful agentic workflows that can run, prompt, and report automatically by looking at a real use case. This video explores a piece of the future of engineering, showing you how to build systems that work for you while you sleep.
🔬 We dive deep into:
• The structure of sequential agentic workflows
• How to use prompt chains for complex reasoning
• Techniques for auto-updating content with AI
• Integrating notifications into your agentic workflows
🚀 Learn to harness the power of:
• Build agentic workflows that can make decisions and take action
• Claude 3.5 Sonnet insane accuracy and instruction following
• Prompt engineering for real-world use cases
👨💻 Perfect for:
• Engineers obsessed with automation
• AI engineers and prompt engineers
• Developers exploring agentic systems
🔑 Key takeaways:
• Step-by-step breakdown of agentic workflow components
• Tips for building robust, self-improving agentic workflows
• Steps to create "living software" that evolves over time
Don't miss this in-depth look at the future of AI development. Like, subscribe, and hit the notification bell to stay updated on the latest in agentic engineering and AI workflows!
🔥 When to use prompt chains:
• When to use Prompt Cha...
✅ Minimalist Prompt Chain Code
gist.github.com/disler/d51d7e...
🔗 Resources
Pub/Sub Notifications: ntfy.sh/
Python Schedule: schedule.readthedocs.io/en/st...
📖 Chapters
00:00 Prompts, Prompt Chains, Agentic Workflows
01:23 Auto Blog Agentic Workflow Demo
02:10 NTFY - Crucial Notification Tool
03:20 Breaking Down the Workflow Steps
05:43 Agentic Step - Running the Prompt Chain
10:30 Sequential Agentic Workflow Steps
14:59 5 Tips for Building Agentic Workflows
16:54 The Future - Less Code, More Agent Work
17:40 LLM Killer User Case - Living Software
#agentic #promptengineer #aiagents - วิทยาศาสตร์และเทคโนโลยี
This is one of the most informative breakdowns of Agentic Workflows I’ve ever seen 😊
As always great video, every Monday you deliver. Please keep with this master class.
Love the animations on your blog!
Great video! It cleared up my confusion about which agent framework to use. Now, I'll ditch them all and stay close to the metal. Your approach resonates with me and gives me the energy to go on with the ai stuff! Thanks a lot for this content.
so glad I found you recently - a subscriber now, and I hope to see more from you in the coming years as we enter this new world of AI
How simple or complicated should the prompts within a system like this be? Are BAPs moot in these cases or can they be a part of it?
EDIT: Ah, there we go, I see. 😅
Great video! I noticed the Gist code is different from the walk through in the video. Is it possible to post the code you used in walk through to get a better understanding on how this agentic process works? Thanks! 👍
Great content!
Great video!
I’m currently in the exploration of giving entire chains like this as tools to an open ai assistant to get a hybrid model that triggers any of the defined work flows based on input.
Oooooo very nice - that is a powerful combination: Give your OpenAI Assistant an array of functions that kick off agentic workflows surrounding your domain / problems.
Hey Dan, thanks so much for the video. One thing I didnt see addressed was the cost of running this? It would seem to me that scraping a website every minute and checking for changes would prove to be extremely costly very fast. Please let me know
Yeah that's a great question. TBH I've been ignoring and eating costs of the agentic workflows I build. When I productionize this in the following weeks it will likely run on some 10 minute interval and I'll experiment with html selectors that focus in on the content of the page that matters to detecting an update.
Important thing to keep in mind, the cost of LLMs is going to zero. Take some risks early and experiment with 'higher cost' agentic workflows so you know what you can do early before the industry catches up to you.
do you know if there is some work of agent networks? like small agents that work together through network/internet, i imagine like dns for agents "find agents for ceirtan tasks domains"
could you share the markdown file :) ?
how can I replace github actions and CI/CD with LLMs?
I am trying to create a github action to do checks, and run makefile pipeline after I merge a PR and neither gpt4o or 3.5 sonnet are able to produce a good solution that works. It's always giving outdated info or errors. So how would that even work if it was automatic and not me even creating and updating manually a yml file for github action.
Do you have a mentoring program or recommended developers for hire?
Not yet. I'm building a resource that will track everything we discuss on the channel and more.
Stay tuned for that.
For potential mentoring/consulting feel free to reach out at agentic@indydevdan.com.
I wait for Monday, as I did years ago for a new episode of Lost
I see your browser keeps popping up, are you scraping from it directly?
Yes
@@indydevdan Can you show us the code scraping a web page directly from browser or the library you used for it?
😎🤖
Exited when I saw 3.5 OPUS... please consider using different background than moving hands.
I'm not sure what you mean. 3.5 Opus comes right from the scrap from anthropic's model page.
I mean now we have 3.5 sonnet only which is already amazing... but 3.5 Opus will be interesting
Yea hands are distracting
Thank you for the useful content. Please remove the overlay style with your hands in the background. It's very distracting.
I find your hands moving in the background extremely distracting. For the rest your content is excellent!
Your audio is always on the lower end.