@tjrobertson52

Why do we still need websites? I'm not saying we don't need websites, but why do we still need them? This is something I've been thinking about a lot lately.

And the more I think about it, the more outdated the website seems to be. I'm not talking about websites like Facebook or TikTok, Google or ChatGPT. For this conversation, let's consider those platforms.

I mean the kind of websites that a business or a brand might have. As far as I can tell, websites only exist for two purposes, to share information and enabling transactions. And then you could include a third, which should just be entity verification.

It's important that you can tell who the information is coming from and with whom you're doing a transaction. So that's important, right? Entities need to be able to share information.

We need to be able to make transactions with entities online and entity being a person, brand or business. What is the website the ideal medium for finding this information and hosting these transactions? Up until recently, search engines were the only tools we had for identifying the entities from whom we wanted to gather information or do transactions with.

But large language models like chat t have already demonstrated how unnecessarily laborious that process is. And I'm not saying that ChatGPT is the ideal interface either. I think AI is still pretty terrible compared to how good it could be.

I just think We've seen enough already to know how much better things could be. Let's take a broad look at what I think is the most general case for websites. You're trying to pick an entity to do a transaction with, so you have some kind of problem and you need a person or organization with a solution.

This could be anything from trying to find someone to hire at your business to trying to find what restaurant to go to for lunch. The current process involves manually looking through pages of content usually created by the brand themselves, and if we're lucky, some third party reviews. But ultimately, we only have time to read a fraction of those.

In other words, we're seeing a small percentage of the information available. It's hard to know how much we can trust the information we're getting. And then it's up to us to figure out how to parse that information and make sense of it so that we can make a rational decision.

Ultimately, what we want is an intelligent system that understands everything that's important to us, can evaluate all the information out there, verify if that information is coming from a trusted source, and return only the context important to us in making the final decision. And I can already hear you saying, but TJ, we can't trust AI to do that.

AI stupid, and you probably already know, I'm gonna say, sure, AI stupid, but it's gonna keep getting smarter, but I actually don't think it needs to get any smarter for the website to be in trouble. Consider the fact that Google's algorithm is still pretty stupid. That doesn't prevent people from using that as their default mode of discovering information.

I actually don't think we need smarter language models to make the website obsolete. We just need a better product. Same Altman of OpenAI has made it pretty clear that they see large language models as a commodity and that they're focusing almost entirely on the product.

Now, of course, it's impossible to say what that product will look like. Otherwise it would already be here. But it's clear that there's a better way.

Whatever that product is, it will need some way of retrieving information, verifying entities, and facilitating transactions. OpenAI, of course, is already trying to do this with ChatGPT, but I think that's just the first step. And currently that's being done through the website.

It's the only place that a brand can host the information and transactions currently. But if we're moving towards a world where people just interface with the product and no longer need to go to the website, the website starts to feel like a really bad medium. We're hosting that information and those transactions.

In the last year, we've seen new protocols coming out like MCP and A2A, which allow agents to communicate with APIs or other agents. The goal of these protocols is to make it as Easy as possible to retrieve information or facilitate transactions. This is the best hint we have at what the future might look like.

My prediction is that each brand will need their own agent or at least an API to verify their identity, share information and facilitate any transactions. My guess is that early on brands will have to build these agents or APIs themselves. But eventually there will be platforms to do it for you.

And the early versions will probably just automate it by tapping into your website. But I still think websites are gonna exist for a long time. They'll probably still be relevant for at least a couple of years.

But I think what your website says about you is gonna matter less and less. What these platforms say about you is going to matter more and more. And your website is just one place that these platforms are getting their information from.

So while we wait to see what will replace the website, if anything, I think the best thing that brands could be doing is just putting their content on as many platforms as possible. Produce high quality content that represents your brand and distribute it across all the social platforms and any other third party websites you can get on to. The goal is to be found everywhere these large language models are looking.