Google Bard, Microsoft ChatGPT, and the Future of Search





Key Information Before We Dive In



What Is LamDA?

Imaging It: Create scenarios and situations based on its understanding of the world.
Talk About It: Conversational chat. Chat ... where have we heard that word recently?
List It: Create lists of tasks and sub-tasks.

What Is PaLM?



"... we’d like to train one model that can not only handle many separate tasks, but also draw upon and combine its existing skills to learn new tasks faster and more effectively.

That way what a model learns by training on one task – say, learning how aerial images can predict the elevation of a landscape – could help it learn another task — say, predicting how flood waters will flow through that terrain.”

What Is ChatGPT?



"People are begging to be disappointed and they will be. The hype is just like... We don’t have an actual AGI and that’s sort of what’s expected of us."



But What's Going On In Search?

What's At Stake For Google And Microsoft?

They can be trained on your styles and mannerisms.
They can be trained to answer emails based on your calendar, in the words you might use.
They can be trained to create complex spreadsheets based on your verbal request vs learning how Excel/Sheets formulas work.
They can be trained to write a paper letter for you, in your style.

But What About SEO?

What happens when the search engine is more confident in their understanding of the world having crawled basically every web page in existence, than they are in you and your content?

What Will The New Search Look Like?





So, Is SEO Dead?

Stable Diffusion Prompt: giant borg stepping over dead human bodies,humans holding laptops and cell phones,in an office,realistic black and white sketch

"It's unlikely that AI-based search engines will make SEOs (search engine optimizers) obsolete. While AI technology is advancing rapidly and has the potential to significantly impact the search engine landscape, SEOs still play an important role in helping businesses optimize their online presence for search engines and user experience."

Then What Do You Mean "Kind Of"?

People will keep buying things.
People will keep seeking solutions to problems.
People will keep looking for entertainment.
People will keep working.



What Skills Will SEOs Need?

Knowledge Graphs - I'm not referring to Google's Knowledge Graph as displayed in search results here, but they're certainly helpful. A Knowledge Graph, as described by IBM, "... represents a network of real-world entities—i.e. objects, events, situations, or concepts—and illustrates the relationship between them. This information is usually stored in a graph database and visualized as a graph structure, prompting the term knowledge “graph.” In short, you can think of them as a graph of concepts and entities (nodes) connecting relationships (edges) that are used to form a "picture" of what an entity is and how it relates to the world at large. IMO, developing your Knowedge Graph is akin to training search engines to understand the relationship your company and/or offerings have to other concepts and things. Want to be included in generated results around SEO? You had best make sure the engines understand fully how you connect to them.
Content - The content I'd be looking at when I'm wanting to "simply rank" and how it's optimized will be fundamentally different than when I'm looking to influence the AI-generated results in a few ways. The first is simply that, given there may be no result or citation, SEOs will need to focus on helping the engines connect the dots with their preferred messaging and topical inclusions (ex: specific products) and less on whether that content might rank for a query that no longer produces a prominent linked result. The second I alluded to earlier, and that's including in your content strategy offsite content on key pages that help form the Knowledge Graph. In this context, it's not links one is seeking, but simply ensuring that your desired associations are understood and reinforced on resources that would be trusted to train a machine learning model. Think, "Wikipedia." But for goodness sake don't spam them.
Content Too Two - On this topic, I also found a comment from Nathan Gotch in his video from February 8 very interesting. In the video he notes not to target simple queries that AI can easily produce an answer for. These are the queries I'm noting to focus on helping them understand that you relate to those queries and should/could be included. He notes that for many queries (like "best XYZ in <my city>") AI cannot fill in the gaps. This is where traditional SEO will still win, so the advice is to target those and I obviously have to agree.
Technical SEO - More important than ever will be technical SEO. Not only in the context of having a fast site to help ensure the bots properly crawl and index what you're offering, but additionally I suspect Schema will play an even larger role as it helps the engines fully understand what an entity is and what it's connected to. Additionally, I suspect we'll be seeing updated APIs and offerings on ways to get your raw content directly to the engines so they can consume it without crawling. This is just a hunch, but I can't see it not happening because crawling is expensive and in this world, unnecessary. You'll still need your XML sitemap, but that too I suspect is due for a revamp and may serve the dual-purpose.

You Have A Bit Of Time ⌛



Some Fun Additional Reading

AI Has Suddenly Evolved to Achieve Theory of Mind

In a stunning development, a neural network now has the intuitive skills of a 9-year-old.

An Introduction to Training LLMs Using Reinforcement Learning From Human Feedback (RLHF)

In this article, we explore Reinforcement Learning from Human Feedback, a novel approach to reducing bias and increasing performance in large language models.

Microsoft Bing's Weird Responses!

Microsoft's new Bing AI can generate some interesting responses when prompted a certain way.

An important next step on our AI journey

The official announcement of Google Bard.