Why is Google scrambling?

Welcome to another edition of what we’re determined to make the best damn newsletter in AI. Here we’ll break down AI topics that matter, open your mind to use cases, and keep you ahead of the curve.

Our #1 goal is to be useful. So please shoot us an email 📩 if you have questions or feedback, and especially if you implement something we share!

Here's what we're covering today:

• Will Search products shift to Answer products?

• A use case focused on helping you learn anything

• This week's update on the AI Wars

Let’s get to it! 👇

TODAY'S PERSPECTIVE

Why is Google scrambling?

If you’re in SEO, digital marketing, or rely heavily on traffic from Google in some way - you’re likely already paying attention to what’s unfolding with the Search Business Model.

But today we’re going to break down why both Google and people who rely on Google are, and probably should be, concerned.

The Search Business Model is part advertising and part content

Google Search is a beautiful product. It helps people find what they are looking for. If you want people to find your things, you need to create helpful content that helps that person. That helpful content makes Google Search better. And in exchange, you get traffic.

If you want to get even more traffic, you can pay (ads). But you even pay less based on how helpful your content is.

The Search Business Model funds Google

In 2021, Google reportedly made ~$150B from search ads alone (58% of its revenue). And if you expand out to their other ads properties, which are tightly connected, ads make up > 80% of Google’s revenue.

The Search Business Model isn’t holding back other players from innovating

In contrast, Bing’s ads business is a tiny portion of Microsoft’s overall revenue. And other "ChatGPT for Search" startups like You.com, Neeva and Perplexity have nothing to lose.

These players can skip the traffic exchange described above, and go straight for answers.

The Search Business Model might be disrupted by the Answer Business Model

The biggest threat to the Search Business Model is that a "successful" Chat-based search experience doesn't end in a link click, or even really look like a search. Most of the time, we just want the answer.

Interfaces like we’re seeing today show how seamless it could be to get answers, backed up by citations and links to learn more. But even these citations function more as credibility, not as your destination. 

The answer is your destination.

The question isn’t what will Google do. It’s what will people do.

If people choose to use an Answer product over a Search product, change is inevitable.

We're digging in more on this with some experts in the space. Stay tuned. 🤓 (and hey, maybe share with a friend while you're at it).

USE CASE DEEP DIVE

Learn anything.

We are very bullish on ChatGPT’s ability to help us learn anything - whether it be via summarization, offering more explanation, or even as simple as having a chat interface over a subject or document that we’re looking to learn.

Today, ChatGPT has 2 major limitations on learning:

1. ChatGPT makes up information, or hallucinates. But will sound confident when doing so. This is because the AI model was trained to generate text, not to be accurate.

2. ChatGPT has a limited context window, or word count that it can process at a time.

One of the best ways around these two issues today is to use a tool that allows you to upload a source document, and generate summaries or ask questions over that document.

These tools allow you to select sections of the document, which not only helps with the word count issue but also gives you the source material to go “fact check” against.

Here are two tools we are using for this.

Explain Paper - upload a PDF and easily highlight any section and ChatGPT will provide an explanation or a summary. You can also ask follow up questions on that section.

Humata.ai - upload a PDF and ask questions over the whole document, with sources.

LINKS

For your reading list 📚

The AI Wars are heat up...

• Meta is joining the AI party, and recently announced LLaMA - their GPT competitor that appears to be better than GPT in some benchmarks

• Snapchat just announced their own ChatGPT, built in close partnership with Open AI

And we're happy to see that ethics are a growing part of the discussion...

• Open AI published how they're thinking about ethics and roadmap in the context of AGI

• We also recently read Futureproof, and found it particularly well-researched in discussing a range of the business ethics and implications of automation (affiliate link)

Member only links (more info on joining here!)

• How to get GPT to sound like you (member hub link, must be signed in!)

• Interested in contributing? Check our #announcements channel for information on how to get involved with The AI Exchange.

That's all!

We'll see you again on Thursday. Thoughts, feedback and questions are much appreciated - respond here or shoot us a note at [email protected]. 

Cheers,

🪄 The AI Exchange Team