The future of web browsing: is it not actually browsing at all?

On the launch of Arc’s next iteration, The Browser Company CEO announced a brand new category of software - “a browser that browses for you”.

In the last few weeks we saw Rabbit launch the R1, a revolutionary new device that uses an LLM to work out the intent behind your instruction, and then an LAM - Large Action Model - to actually do that thing for you. The limitation of LLMs, that they can’t actually do anything, only tell you or show you things, is surpassed with a Large Action Model.

If we can accomplish this, our interaction with machines will take a huge leap forward again. Now, we' still require human intervention to use the output of LLMs in a way that makes sense in our context. With LAMs, we don’t have that holding us back.

Browsing the web without browsing

When we search the internet, we find information through researching files. We dig through public pages that others have created to locate the stuff we need.

It requires thought, intuition, and occasionally, perseverance. Search engines do a good job at surfacing the information that they think we need, as quickly as possible. But we are still limited but our own biases and visual searching patterns to find that information.

Arc can now search the web and bring information back, to display it all in one page, customised to your query. It eliminates the to-and-fro - or the ‘pogoing’ - that you might experience when searching for your results.

It’s not perfect. It still assumes that the machine’s processing of the information is correct and that it understands our query, and is providing information that’s accurate and sensible. We know that LLMs often hallucinate or provide completely wrong answers.

But it feels like we’re on the precipice of something pretty revolutionary in the way that we use the internet.

The elimination of research

Our use of browsers and the traditional hunter-gatherer approach to explore the internet could be a thing of the past, and faster than we might think. Removing the need for reviewing search results, digging through ad-rammed pages and chumboxes, and having to identify useful content quickly, could spell a totally new way of using the technology around us.

Our Alexas and Homepods have been deployed in our homes for years now, and with Siri and Hey Google in our pocket, technologists have been slowly training us to communicate with the internet in a way that’s faster, more call-and-response, and moves in the direction of trusting the response that is given to us by the machines. Rather than asking for 10 results when we speak to Alexa, we’re pretty content to take the first answer as red. We’re already happy with the first result.

Doing the same thing with the internet feels like a natural evolution at this point.

Dying creativity of the web

But what does this mean for the internet? The creativity we pour into responsive, dynamic, beautiful webpages and systems designed to capture and maintain attention are created for the human eye, balanced with the requirement to be read accurately by a search engine robot. Will we need to design user interfaces for humans in the future, when our interactions might end up being reduced to a voice-first interface or a call-and-response directly through our browsers?

What will our shopping experiences look like when we can ask a machine to show us exactly what we want, and it creates a personalised experience for us every time - in exactly the way we want it?

Are information sites, travel guides, review websites and news pages going to be relevant in five years time when all their content is simply processed, reconstituted and delivered to the viewer devoid of original creativity and context?

I’m really not sure how I feel about things at this point. We could be about to see a fundamental shift in the way the internet is not just accessed, but also created and maintained by brands, organisations and individual contributors.

I wouldn’t say I’m worried exactly, but nervous. LLMs have already had a seismic shift in the way we use technology, and have had that impact far faster than we could have imagined when GPT3 was released. But I fear we aren’t prepared for what is around the corner.