Almost Timely News, March 24, 2024: A Guided Tour of the Most Powerful Generative AI Model Google Offers

Almost Timely News: A Guided Tour of the Most Powerful Generative AI Model Google Offers (2024-03-24) :: View in Browser

Almost Timely News

Check out these two new talks, free for you to enjoy:

Content Authenticity Statement

90% of this week’s newsletter was generated by me, the human. You will see many outputs of generative AI in the opening section. Learn why this kind of disclosure is a good idea and might be required for anyone doing business in any capacity with the EU in the near future.

Watch This Newsletter On YouTube πŸ“Ί

Almost Timely News: A Guided Tour of the Most Powerful Generative AI Model Google Offers

Click here for the video πŸ“Ί version of this newsletter on YouTube Β»

Click here for an MP3 audio 🎧 only version »

What’s On My Mind: A Guided Tour of the Most Powerful Generative AI Model Google Offers

This week, we’re going to go on a guided tour of Google’s AI Studio and the most powerful generative AI model Google has to offer, Gemini 1.5 Pro. Now, before we begin, we should clarify a few things. Google, in its fine tradition of naming things as confusingly as possible (those of you who work in Analytics, you know the joy of Google Tag Manager and the new Google Tag which is NOT part of Tag Manager despite the name) has the Gemini models themselves, but also a web interface named Gemini. Today, we’ll be looking at the model, not the interface.

To access this, you’ll go into Google’s AI Studio, which is also confusing because it’s not the Google Vertex AI Studio in Google Cloud. Google AI Studio is for prototyping stuff that you’ll then later deploy elsewhere. Yeah, don’t blame me, I didn’t name any of this stuff. You obviously need a Google account, and to put your stuff in production, you’ll need a Google Cloud account with billing turned on.

But for just casual use, this is a great place to start. For the demo use case, we’ll be constructing a guide for growing an Instagram account. First, let’s familiarize ourselves with the interface, which is probably my favorite for working with generative AI so far. It’s not as cluttered as OpenAI’s developer portal, but it’s clean enough that a non-technical user can use it easily.

I’ll call your attention to six key features, noted on the screenshot below.

Google AI Studio interface

  1. This is where you create new prompts to test. Google offers three kinds – chat, freeform (aka autocomplete), and structured. Most people will use chat. If you’re comfortable with ChatGPT and similar systems, choose chat.
  2. This is where your saved work will show up. This relies on Google Drive and your interactions are saved to Google Drive.
  3. This is where you choose the model you want to use. We’ll go with Gemini 1.5 Pro, which is their latest and greatest.
  4. This is where you can control the safety settings. In testing, this matters less, but in production, you’ll want to make sure these are set to your use case.
  5. This is where you connect to assets like images, videos, files, and folders. This includes Google Drive as well as uploads – which is a lot more than the Gemini web interface offers.
  6. This is the context window, arguably the most important part of this entire thing. This tells you how close to full the short-term memory of the model is, and once it’s full, it will start forgetting the earliest part of your conversation. This resets per conversation, so if you start a new prompt session, this goes back to zero.

The last part, the context window, is 1 MILLION tokens in Gemini 1.5 Pro. It’s difficult to express just how important this is. The context window is a model’s short-term memory, what it knows during your conversation. When we start a session with any AI tool, this is empty. As we give it prompts and interact with it, the short-term memory fills up. Over time, it gets full and when it does, the model forgets the earliest content. If you’re used to the free version of tools like ChatGPT, you know this forgetting happens very, very quickly.

A million-token context window is about 700,000 words. That’s 10 good-sized business books – that’s how much this model can remember in the short term. That’s incredible, because it means we can have it work with very large pieces of content, have extended conversations, and have it remember things we said quite some time ago. We can give it lots of reference material to think through and still not fill up the memory.

Let’s put this into action. Earlier today, my martial arts teacher, Mark Davis of the Boston Martial Arts Center, asked me for some help growing our school’s Instagram account (go follow them if you like). Now, I am not an Instagram expert. One glance at my own Instagram account makes it clear it’s not my specialty. But there are lots of people who share lots of ideas on the topic. If we could gather up the data that exists publicly, distill it, synthesize it, and summarize it, could we come up with a highly effective set of strategies and tactics?

Here’s what I did: I found every podcast on the topic of growing your Instagram account in 2024 on YouTube and fed them all to OpenAI’s Whisper transcription software, run locally. This costs nothing because it uses your computer to do the processing. Now I’ve got a folder full of podcast transcripts. (I suppose I could have just gone to the podcasts’ sites themselves for the transcripts, but it’s easier to just let AI do this).

I’ll take each transcript and load it into Gemini 1.5, asking it to summarize the transcript this way:

You are an accomplished professor and researcher at MIT. You sit on review boards, referee panels, and peer review committees. You review academic papers and publications.

One of your skills is a technique called paper to protocol, in which you digest a paper’s findings and translate those findings into a protocol, a series of actions an individual can take based on the research.

For example, if a peer-reviewed paper finds that reducing sodium intake helps improve health, you would create a protocol that says to limit your sodium intake.

Or, if a peer-reviewed paper finds that using adverbs less in advertising copy is more effective at driving sales, you would recommend a checklist for removing adverbs and replacing those adverbs with another part of speech.

Protocols should be firmly grounded in the research, backed up by statistically meaningful conclusions from the research.

Protocols should be action-oriented; they should provide clear, unambiguous directions about what someone should do or change in their behaviors to achieve results aligned with the findings of the paper.

Today, you’re going to read through a transcript discussing best practices and research around Instagram and growing Instagram accounts. You’ll summarize the information you read into a set of protocols. First, you’ll make an outline of the key findings as presented in the transcript. Then, of those findings, you’ll distill them down to protocols.

Here’s the format to use:

“””
Finding Title

Findings:
Finding 1
Finding 2

Protocols:
Action steps to take 1
Action steps to take 2
“””

Here’s a quick look at the process as Gemini chews through each transcript:

Gemini output

After Gemini processes all the transcripts, I asked it to do the following:

Great. Now that you have consumed all these transcripts, categorize all the advice they give and restructure all the advice and protocols into one master outline.

And here’s what Gemini spit out, distilling down all that information:

Gemini Screenshot

Now here’s the thing: if you know Meta, and you know their transparency center, they talk about how their systems work at a high level. Lots of people talk about the Instagram algorithm, but very few seem to know how the system overall works, even though Meta has told us. If we add Meta’s documentation to the mix and tell Gemini to resolve conflicting advice by leaning in favor of what Meta has told us about their own system, we get a comprehensive list of recommendations that’s more balanced and aligned with how Meta says their systems work.


Master Outline for Instagram Success in 2024: Synthesizing Expert Advice and Meta Documentation

This master outline combines insights from expert advice in the transcripts with official documentation from Meta to provide a comprehensive and actionable Instagram strategy for marketers in 2024. In cases of conflicting information, Meta’s documentation is prioritized as the source of truth.

I. Account Setup and Optimization
– Define Your Niche:
– Identify your ideal audience: demographics, interests, challenges.
– Determine your specific topic within your broader niche.
– Refine your niche over time as you learn and grow.
– Optimize Your Profile:
– Username: Choose a username that reflects your brand and incorporates relevant keywords if possible.
– Name: Use your business name or a keyword-rich description, different from your username.
– Profile Photo: Choose a high-resolution, recognizable image that fits well within a circular crop.
– Bio: Craft a concise and impactful bio that highlights key aspects of your business, uses a voice and style that resonates with your target audience, and includes a clear call to action.
– Website Link: Use a trackable link shortener and customize it to reflect your brand or campaign.
– Upgrade to a Business Profile: Unlock valuable features like contact buttons, analytics, and advertising options.

II. Content Strategy and Creation
– Content Mix:
– Utilize a mix of Reels, carousel posts, stories, and potentially broadcast channels to maximize reach and engagement across different Instagram surfaces.
– Experiment with different posting frequencies and content formats to find what works best for your audience.
– Content Types:
– Reels: Create a mix of original and trendy Reels. Plan original Reel concepts in advance and leverage trending audios strategically.
– Carousels: Share photo dumps, behind-the-scenes moments, and other engaging content to connect with your community.
– Stories: Post consistently, allowing older stories to expire for optimal reach. Focus on creating high-quality stories with engaging visuals and concise text.
– Broadcast Channels: Share exclusive content and updates, host Q&A sessions, and build deeper connections with your most engaged followers.
– Content Principles:
– Focus on Share-worthy Content: Create content that evokes emotions, inspires action, is relatable, or provides breaking news within your niche.
– Provide Value: Offer entertainment, education, motivation, or inspiration to keep your audience engaged and coming back for more.
– Prioritize Quality over Quantity: Invest time and effort in creating high-quality content that resonates with your audience, rather than just churning out content for the sake of posting.

III. Engagement and Growth Strategies
– Understand the Algorithm: Stay informed about how the Instagram algorithm works and the various factors that influence content ranking and reach. Adapt your strategy based on the latest updates and insights.
– Optimize Posts for Discovery: Utilize location tags, keywords in captions and alt text, and relevant topics to help the algorithm categorize your content and deliver it to the right audience.
– Engage with Your Audience Proactively: Dedicate time daily to interact with your niche community, larger brand accounts, and your current followers. Build relationships and signal relevance to the algorithm.
– Study Your Competitors: Analyze their content and strategies to gain valuable insights and inspiration for your own approach.
– Collaborate with Influencers and Relevant Accounts: Partner with influencers and other accounts in your niche to expand your reach and attract new followers.
– Leverage Other Platforms: Promote your Instagram account on platforms like Pinterest and YouTube to drive traffic and growth.
– Analyze Your Insights and Replicate Success: Regularly review your Instagram insights to identify what content performs best and adapt your strategy accordingly.
– Be Consistent and Patient: Develop a sustainable posting schedule and stick to it. Understand that organic growth takes time and effort.

IV. Additional Considerations
– Experiment with different content formats: Consider the impact of videos, photos, and carousels on the algorithm and experiment to see what resonates best with your audience.
– Utilize Instagram’s built-in features: Explore and leverage Instagram’s native features, including effects, filters, and music options, to enhance your content and potentially gain an algorithmic boost.
– Stay informed about platform updates: Keep up with the latest changes and updates from Meta regarding the Instagram algorithm and adapt your strategy accordingly.


This is helpful. Here’s the important part in this experiment. Take a look at the bottom of Google’s AI Studio:

AI Studio

I have consumed 76,716 tokens in this chat. That exceeds what the consumer version of Gemini can handle by about double. It’s more than halfway to what the paid version of ChatGPT in the consumer interface can handle. And it’s 7.3% of what Gemini 1.5 can handle. If I wanted to, I could go grab a hundred more podcast episodes or YouTube videos about Instagram strategy and add them into the short-term memory, coming up with the absolute ultimate guide to Instagram – or any other topic.

That’s the power of a huge context window, of a huge short-term memory. We can take the power of large language models along with vast amounts of data we provide and turn that data into useful insights and action steps to be taken.

Now, after all this, I’m still not an Instagram expert. The synthesized advice that Gemini helped me to assemble is still only advice. I or my teacher still need to do it, to take action on it, and to bring it to life. That’s the hard part – advice is easy, execution is difficult. But in terms of getting advice in the first place, this makes things much faster.

Let’s briefly talk about what we didn’t do in this experiment. I didn’t spend a lot of time vetting the podcast episodes or transcripts. I went off popularity, which episodes in 2024 got the most attention. Popularity doesn’t mean correctness, so if you’re using this strategy for high-value, high-risk endeavors, you’ll want to curate the source data better.

I didn’t run this by a human subject matter expert. As a marketer, the advice looks generally correct, but if this were a high-stakes engagement, I’d want to run it past a real Instagram subject matter expert for their take on it – especially what’s missing.

And finally, I didn’t evaluate my teacher’s actual Instagram profile against this list of best practices, because this is a newsletter about using a large language model with a very large context window, not about Instagram best practices. I’ll be sending him the list of best practices, and this video walkthrough, but we would spend a lot more time with each of the best practices than we did.

Now, I turn this over to you. Take this advice and run with it in your industry, at your company, with your content. Get a comprehensive corpus of top content about any given subject and use a language model to consolidate, interpret, and distill it down into something you can take action on. Convert information into action, and reap the benefits of taking action on all the advice you’ve consumed.

How Was This Issue?

Rate this week’s newsletter issue with a single click. Your feedback over time helps me figure out what content to create for you.

Here’s The Unsubscribe

It took me a while to find a convenient way to link it up, but here’s how to get to the unsubscribe.

Click me to unsubscribe!

If you don’t see anything, here’s the text link to copy and paste:

https://almosttimely.substack.com/action/disable_email

Share With a Friend or Colleague

If you enjoy this newsletter and want to share it with a friend/colleague, please do. Send this URL to your friend/colleague:

https://www.christopherspenn.com/newsletter

For enrolled subscribers on Substack, there are referral rewards if you refer 100, 200, or 300 other readers. Visit the Leaderboard here.

ICYMI: In Case You Missed it

Besides the new Generative AI for Marketers course I’m relentlessly flogging, I did a piece this week on the two things you’ll need to truly succeed in the long-term with generative AI.

Skill Up With Classes

These are just a few of the classes I have available over at the Trust Insights website that you can take.

Premium

Free

Advertisement: Generative AI Workshops & Courses

Imagine a world where your marketing strategies are supercharged by the most cutting-edge technology available – Generative AI. Generative AI has the potential to save you incredible amounts of time and money, and you have the opportunity to be at the forefront. Get up to speed on using generative AI in your business in a thoughtful way with Trust Insights’ new offering, Generative AI for Marketers, which comes in two flavors, workshops and a course.

Workshops: Offer the Generative AI for Marketers half and full day workshops at your company. These hands-on sessions are packed with exercises, resources and practical tips that you can implement immediately.

πŸ‘‰ Click/tap here to book a workshop

Course: We’ve turned our most popular full-day workshop into a self-paced course. The Generative AI for Marketers online course is now available and just updated this week! Use discount code ALMOSTTIMELY for $50 off the course tuition.

πŸ‘‰ Click/tap here to pre-register for the course

If you work at a company or organization that wants to do bulk licensing, let me know!

Get Back to Work

Folks who post jobs in the free Analytics for Marketers Slack community may have those jobs shared here, too. If you’re looking for work, check out these recent open positions, and check out the Slack group for the comprehensive list.

Advertisement: Business Cameos

If you’re familiar with the Cameo system – where people hire well-known folks for short video clips – then you’ll totally get Thinkers One. Created by my friend Mitch Joel, Thinkers One lets you connect with the biggest thinkers for short videos on topics you care about. I’ve got a whole slew of Thinkers One Cameo-style topics for video clips you can use at internal company meetings, events, or even just for yourself. Want me to tell your boss that you need to be paying attention to generative AI right now?

πŸ“Ί Pop on by my Thinkers One page today and grab a video now.

How to Stay in Touch

Let’s make sure we’re connected in the places it suits you best. Here’s where you can find different content:

Advertisement: Ukraine πŸ‡ΊπŸ‡¦ Humanitarian Fund

The war to free Ukraine continues. If you’d like to support humanitarian efforts in Ukraine, the Ukrainian government has set up a special portal, United24, to help make contributing easy. The effort to free Ukraine from Russia’s illegal invasion needs your ongoing support.

πŸ‘‰ Donate today to the Ukraine Humanitarian Relief Fund Β»

Events I’ll Be At

Here’s where I’m speaking and attending. Say hi if you’re at an event also:

  • Society for Marketing Professional Services, Boston, April 2024
  • Lab Products Association, Nashville, April 2024
  • Society for Marketing Professional Services, Los Angeles, May 2024
  • Australian Food and Grocery Council, Melbourne, May 2024
  • MAICON, Cleveland, September 2024
  • MarketingProfs B2B Forum, Boston, October 2024

Events marked with a physical location may become virtual if conditions and safety warrant it.

If you’re an event organizer, let me help your event shine. Visit my speaking page for more details.

Can’t be at an event? Stop by my private Slack group instead, Analytics for Marketers.

Required Disclosures

Events with links have purchased sponsorships in this newsletter and as a result, I receive direct financial compensation for promoting them.

Advertisements in this newsletter have paid to be promoted, and as a result, I receive direct financial compensation for promoting them.

My company, Trust Insights, maintains business partnerships with companies including, but not limited to, IBM, Cisco Systems, Amazon, Talkwalker, MarketingProfs, MarketMuse, Agorapulse, Hubspot, Informa, Demandbase, The Marketing AI Institute, and others. While links shared from partners are not explicit endorsements, nor do they directly financially benefit Trust Insights, a commercial relationship exists for which Trust Insights may receive indirect financial benefit, and thus I may receive indirect financial benefit from them as well.

Thank You

Thanks for subscribing and reading this far. I appreciate it. As always, thank you for your support, your attention, and your kindness.

See you next week,

Christopher S. Penn


You might also enjoy:


Want to read more like this from Christopher Penn? Get updates here:

subscribe to my newsletter here


AI for Marketers Book
Take my Generative AI for Marketers course!

Analytics for Marketers Discussion Group
Join my Analytics for Marketers Slack Group!



Comments

Leave a Reply

Your email address will not be published. Required fields are marked *

Pin It on Pinterest

Shares
Share This