AI Agent Insider Edition #005 - How to Win a DeLorean with Agentic AI + Other Exciting News & Resources

👋🏻 Hello and welcome to AI Agent Insider!

Where I round up the best news and learning resources on AI agents and assistants that dropped this week.

What’s in it for you?

• You become more productive with GenAI-powered workflow automation, every Saturday, for free.

• You save hours of time by not having to pan through the mountain of AI-related content produced each week to find the most practical, actionable (Agentic AI) gold.

• You’ll learn at least one useful thing on agentic AI use cases, tooling, and techniques, if you commit to reading/trying one or more of the resources shared in AI Agent Insider each week.

All of this will add up over time to help you automate, execute, and create value like a workflow automation boss.

Not to mention the edge you could gain with your up-to-date knowledge of the most practical, in-demand AI skillsets that one should be learning and applying right now.

Sound good? Then let’s dive in!

In today’s issue:

• AI CARTOON OF THE WEEK

• AI AGENT TWEET OF THE WEEK

• THE RESOURCE ROUNDUP

• AI MEME OF THE WEEK

Not a subscriber yet? Sign up free here.

AI CARTOON OF THE WEEK
Career Counseling...

AI AGENT TWEET OF THE WEEK
Low-code tutorial in the thread.
Give it a whirl!

BEST LINKS THIS WEEK
The Resource Roundup
(Read, Watch, Do)

READ:

• Anthropic’s AI now lets you create bots to work for you
  _{Source: The Verge}

  This week, Anthropic announced that they’ve made a new function calling feature “generally available across the entire Claude 3 model family through the Anthropic Messages API, Amazon Bedrock, and Google Vertex AI.”
  
  They’re calling this “Tool Use”. You can learn more by diving into the docs to try it out. There is also a free course that Anthropic has put together, which you can check out here (WARNING: if Github scares you as much as the term “function calling”, then this tutorial might not be for you… However, I’ll be sharing related no-code/low-code tutorials on it moving forward, so stay tuned).

  
  

• Microsoft, Beihang release MoRA, an efficient LLM fine-tuning technique
  _{Source: VentureBeat}
  Last week, we shared a resource on LoRA. This week, meet MoRA.
  Mora to come on both techniques in future editions 😉.

• HATE TO BREAK IT TO YOU BUT NEW RESEARCH SUGGESTS:
  Fine-Tuning LLMs Breaks Their Safety and Security Alignment
  _{Source: Robust Intelligence Blog}
  This article dives into and explains the recent security research conducted by Robust Intelligence, which should be eye opening, especially for the many enterprises who are already working on customizing foundational models for their business.
  
  In their testing, they observed the following:

• INTERESTING PERSPECTIVE: Hi, AI: Our Thesis on AI Voice Agents | Andreessen Horowitz
  _{Source: a16z Blog}
  

• BUT WAIT, THERE’S MORE:

  • Arista announces AI networking agent, Nvidia and Vast partnerships
    _{Source: NETWORKWORLD}
    

  • Opera is drafting in Google Gemini to help supercharge the browser’s Aria AI assistant
    _{Source:TechRadar}
    

  • This week’s LangChain Release Notes, highlights which include their recently improved documentation structure and content. There are now dedicated sections that make it easy to dive into their tutorials and how-to-guides.
    _{Source: LangChain Blog}
    

WATCH:

• SPICY TAKE THAT I LOVE: Have You Picked the Wrong AI Agent Framework?
  _{Source: Matt Williams (Youtube)}
  Emphasizes the importance of not being so in love with Agentic AI that you lose sight of choosing the best tool for the job to be done (i.e., Agents might be overkill! Don’t use a fancy Dremel to remove a nail, when a simple hammer makes more sense to use…)

• CODE ALONG: Building News AI Agents Using CrewAI And Google Gemini Pro LLM Models
  _{Source: Krish Naik (Youtube)}

  • The sample code from the tutorial is linked here.

DO:

• Dive into the epic prompt library that Anthropic provides for free in their documentation. I didn’t realize this gem existed!

• Start building with the Anthropic API, so that you can get familiar with and try out the “tool use” integration capabilities that Anthropic just rolled out to Claude models (see above).

• Consider registering for and participating in the Google Gemini API competition that is running from May 14th - Aug 12th.

  • The grand prize is a custom electric 1981 DeLorean!!!

  • Even if you don’t win the DeLorean, you can still win big bucks and learn a ton.

  • But man… A custom electric DeLorean would be pretty sweet.
    
    If you want to form a team to compete this summer… Lmk!

• Consider registering for and participating for the Microsoft Learn Challenge: Build Edition that is running from May 21st – June 21st, 2024.

AI MEME OF THE WEEK
Conversation starter to avoid…

SHARE AI AGENT INSIDER
Know someone who would love this newsletter?

As my 4.5 yr old twins love to say, “sharing is caring!”

Plus, swag and resources are on the way for those who share!

Consider sharing this with folks in your network who just might love this newsletter, and I’ll ensure that you, dear founding reader, are the first to receive the foreshadowed swag.

THE GOOD, THE BAD, THE UGLY…
Bueno? Meh? Feedback Welcome!

If you took the time read to this section, that means a lot to me.

Seriously, if there’s anything in particular you would like for me to cover in my next editions, please reply to this email and let know.
 👇 🙏