The TLDR

Highlight: We’ve worked on many AI projects over the past few months. Today, we’re focusing on helping you “glue” these projects together using a tool called Zapier, which will help your systems work for you while you sleep 😴

Musing of the Week: Can we lean on the U.S. government for leadership in AI regulation? We dig into Biden’s recent executive order, which seeks to regulate the development and use of AI.

⚡️You should be Zapping everything

What happens when you multiply creation x automation? 🤯 Today we’re experimenting with combining AI’s ability to create with Zapier’s ability to automate.

Our readers in tech have probably heard of Zapier, the workflow automation company that made headlines for basically bootstrapping to a $5 billion valuation (the company took just $1.3M in initial funding before swearing off of VC money).

🥸 For our friends who haven’t heard of Zapier before, here’s an overview:

• Zapier is a workflow automation company that provides users with the ability to connect the different services/tech that they use.

• For example - if you use SurveyMonkey to capture customer leads and Gmail to reach out, you could set up an automation to send an automatic outreach email to each lead, rather than waiting until you have time to do it yourself.

  • SurveyMonkey and Gmail don’t naturally have any way to “talk” to each other, but Zapier enables itself as the middleman.

• Another example - maybe you want to automatically add “meeting prep” to your to-do list on Asana every time a new client meeting pops up on your Google Calendar. Zapier can help you do just that.

  • Same thing - Google Calendar and Asana don’t have any way to “talk” to each other, but Zapier allows the Google Calendar to pass along information.

• Zapier is super user-friendly, and you don’t need to code at all.

• Each automation you build in Zapier is called a “zap.”

• Professionals everywhere, from Etsy shop owners to corporate consultants, are using Zapier to save 10+ hours weekly.

• Zapier has a free plan (1 Zap per workflow, 100 tasks/month) as well as robust plans (more Zaps per workflow, more tasks/month)

So, that kind of automation sounds pretty powerful. What happens when you layer AI on top of that?

We’re in the middle of a new project that requires a ton of cold email outreach, so we tested out a ChatGPT + Zapier workflow for that task. Happy to report it successfully saved us several minutes per email. The Zap can be used indefinitely, so the 30 minutes of build time was super worth it.

Here’s our 3-step workflow with 2 Zaps:

1. Set up your tech tools

   • Create a new Google Sheet to input your prospect information (make sure to name Header columns properly)

   • Create a Zapier account and log in

2. In the Zapier dashboard, hit “Create” → “New Zap”

   • Select the Trigger (which is Zapier-speak for the event that kicks off the workflow). Our Trigger is “Trigger when a new row is added to Google Sheet”

     • You will be prompted to connect to your Google Account and select your Google Sheet name

   • Add the second action, which is “Send the prospect details to ChatGPT for generating a personalized email”

     • You will be prompted to your ChatGPT account so that Zapier can use it

     • In Zapier, provide a cold outreach template that ChatGPT can use to generate emails. Use the data types from your prospect sheet (e.g., Name, Business Type) to create fields in the template so that ChatGPT can customize.

   • Add the third action, which is to “Save the generated email as a draft in Gmail”

     • You will be prompted to connect to your Gmail inbox

3. Add a few test prospects to your Google Sheet, give the workflow 4-5 minutes to process, and check for the results in your Drafts. Tweak your Zaps and prompts as needed! Hope you saved a few minutes.

P.S. Zapier now has an AI assistant that can build your workflows for you, but we provided the manual steps so you can take a look at each step as you build.

🧠 Musing of the Week

On Monday, President Biden took a major step by issuing an executive order aimed at regulating AI. If you're a regular reader of The Multidisciplined, you know we've been grappling with the questions surrounding AI regulation for some time. We're both excited and apprehensive—AI has the potential to improve many aspects of our lives, but it also poses complex ethical and existential risks.

Biden's executive action attempts to strike a middle path. It's not a laissez-faire approach allowing AI to evolve unchecked, and it doesn’t seem to be a cumbersome regulatory burden that stifles innovation. For now it mainly targets the biggest players like OpenAI and Google. These giants will be required to share safety tests for their largest models, specifically those exceeding 100 septillion floating point calculations (that’s a 1 followed by 24 zeroes). This nuance is crucial, as it should shield startups from the brunt of regulatory burdens that could otherwise provide a competitive advantage to existing large companies.

This order comes almost a full year after the release of ChatGPT-3.5, which kicked off the AI arms race last November. While it's encouraging to see regulatory wheels turning, we need to ask: can governance keep pace with the frenetic speed of AI advancements? Since the potential of AI is growing more by leaps in hardware and software than by changes in consumer behavior, the pace of development will likely only increase. The real question is whether these efforts can evolve swiftly enough to effectively manage a technology that's advancing at breakneck speed. Regardless, from where we’re sitting, this seems like a positive start.

The executive order is fairly extensive, but here’s a few key takeaways:

1. Share Safety Test Results: Require developers of high-risk AI systems to share safety test results with the U.S. government.

2. NIST Standards: The National Institute of Standards and Technology (NIST) will set rigorous standards for AI safety and security.

3. Privacy Legislation: The President is calling on Congress to pass bipartisan data privacy legislation.

4. Combat AI-Enabled Fraud: Develop standards for detecting AI-generated fraudulent content.

5. Address Algorithmic Discrimination: Provide guidelines to federal departments on best practices for investigating and prosecuting civil rights violations related to AI.

6. Responsible AI in Healthcare: Promote the safe and effective use of AI in healthcare, including a safety program to report harm or unsafe practices.

7. AI's Impact on Workers: Develop best practices to address job displacement, labor standards, and workplace equity in an AI-driven labor market.

8. National AI Research Resource: Launch a national resource to provide AI researchers and students with key resources and data.

9. Global AI Collaboration: Engage in international collaborations to establish robust frameworks for AI's safe and ethical use.

10. Federal Agency Guidelines: Issue guidance for how federal agencies should use AI, including standards for rights, safety, and procurement procedures.

🙌 If you’re hyped about the generative AI industry specifically, here are some of the coolest roles we’ve seen this week:

• Sales Operations Analyst, Generative AI @ Amazon

• Product Marketing Manager @ Gleen AI

• Artificial Intelligence, Business Development, CPG @ Deloitte

• Operations Manager, Public Sector @ Scale AI

🔨 Check out these other AI tools we’ve been looking at this week:

• Sonara.ai - Job search automation platform that applies for you, until you find a job

• Unscreen - Intelligently and automatically removes your video background. No more green screens for you!

• TLDR This - For when you just need the gist, TLDR This summarizes any piece of text content

That’s all for this week. See you next Tuesday!

Lorel & Reily