Richard Lawrence

A data, tech & SEO nerd who runs a bit & does some other stuff occasionally. Principal - Data Science, Performance Marketing & SEO @sanity_io

[Richard Lawrence @richlawre]

Just in case anyone is unaware, you can access Wikipedia analytics. Can be very useful as an extra data point vs. Google Trends & helps understand intent.

https://t.co/lEqR4w1DJO

[Richard Lawrence @richlawre]

@DanielNotDan21 @danbarker @danbarker seems like one of the least biased people I follow. A genuine truth-seeker.

[Richard Lawrence @richlawre]

@nickwilsdon Same name as Google's new MMM 🤷‍♂️ some irony in there somewhere.

[Richard Lawrence @richlawre]

SEO is no longer just about optimizing content on your website.

It's about structuring & storing content so you can surface it for LLMs to learn deeply about who you are and what you do.

Find out more about how Sanity's Content Lake can help: https://t.co/D55RZaBV6e

[Richard Lawrence @richlawre]

SEO in the AI era involves a lot more than just web content.

I've seen a lot of tactics being spoken about in terms of SEO for LLMs (or whatever you want to call it!)

Whilst it's important to consider whether these are right for your brand, you also need to think longer term.

[Richard Lawrence @richlawre]

@lilyraynyc Also when you search for a business address, giving outdated info even though it's correct in Google Maps.

[Richard Lawrence @richlawre]

@simeonGriggs @RealDMitchell on Fridays for a treat

[Richard Lawrence @richlawre]

Anyone else been so rude to their LLM that its tone has permanently changed?

[Santiago @svpino]

The Model Context Protocol (MCP) is not just "another API lookalike." If you think, "Bro, these two ideas are the same," it means you still don't get it.

Let's start with a traditional API:

An API exposes its functionality using a set of fixed and predefined endpoints. For example, /products, /orders, /invoices.

If you want to add new capabilities to an API, you must create a new endpoint or modify an existing one. Any client that requires this new capability will also need modifications to accommodate the changes.

That issue alone is a colossal nightmare, but there's more.

Let's say you need to change the number of parameters required for one endpoint. You can't make this change without breaking every client that uses your API! This problem brought us "versioning" in APIs, and anyone who's built one knows how painful this is to maintain.

Documentation is another issue. If you are building a client to consume an API, you need to find its documentation, which is separate from the API itself (and sometimes nonexistent.)

MCP works very differently:

First, an MCP server will expose its capabilities as "tools" with semantic descriptions. This is important! Every tool is self-describing and includes information about what the tool does, the meaning of each parameter, expected outputs, and constraints and limitations.

You don't need separate documentation because the interface itself is that documentation!

One of my favorite parts is when you need to make changes:

Let's say you change the number of parameters required by one of the tools in your server. Contrary to the API world, with MCP, you won't break any clients using your server. They will adapt dynamically to the changes!

If you add a new tool, you don't need to modify the clients either. They will discover the tool automatically and start using it when appropriate!

But this is just the beginning of the fun:

You can set your tools so they are available based on context. For example, an MCP server can expose a tool to send messages only to those clients who have logged in first.

There's a ton more, but I don't think I need to keep beating this dead horse.

AI + MCP > AI + API

*micdrop*

[Richard Lawrence @richlawre]

@kmelve I think my most common is 'final try' countless times in a row over a number of hours whilst hating myself for not being able to walk away.

[Richard Lawrence @richlawre]

@PeteGould_ Well we were both wrong. Robert Peston it is. https://t.co/0SJoHdIuUC

[Richard Lawrence @richlawre]

@PeteGould_ Paul Chuckle

[Richard Lawrence @richlawre]

@kmelve New Noise was always the first song to get us out of our seats at our favourite student union night

[Richard Lawrence @richlawre]

* I used the Wikipedia API to get this data, the most popular pages endpoint.

[Richard Lawrence @richlawre]

Did you know the most viewed Wikipedia page in 2024 was Cleopatra? More than the US election, Trump or Kamala Harris?

Couldn't understand it, then stumbled on the answer - Google Assistant prompts you to 'Show Cleopatra on Wikipedia'. Surprising it makes that much difference.

[Richard Lawrence @richlawre]

Here's the prompt on Google Assistant. Never actually paid attention to these!

[Richard Lawrence @richlawre]

How did LinkedIn grow to the size it is when it's held together by tape? Trying to refresh access_token, says I can't because the state parameter was modified. Solution is maybe to log out & in. Try to login & asks for a code to be sent to my phone which never arrive. The end.

[Richard Lawrence @richlawre]

Well this is a bit annoying

[Richard Lawrence @richlawre]

@jamesfgibbons The first time I used it, it seemed to be better...was clearer and more specific about the problem and solution. GPT-4o is really poor at being direct. But Claude Sonnet 3.5 was still wrong, regardless. And same for each time since.

[Richard Lawrence @richlawre]

Don't really get the hype around Claude Sonnet 3.5. It hasn't managed to help me with anything that GPT-4o hasn't been able to help me with. It's also been incorrect in each case so far.

[OpenAI @OpenAI]

We’re testing SearchGPT, a temporary prototype of new AI search features that give you fast and timely answers with clear and relevant sources.

We’re launching with a small group of users for feedback and plan to integrate the experience into ChatGPT. https://t.co/dRRnxXVlGh