Didn’t even think of it as a possibility. WTF would a browser need with LLM?
I’d love a browser-embedded LLM that had access to the DOM.
“Highlight all passages that talk about yadda yadda. Remove all other content. Convert the dates to the ISO standard. Put them on a number line chart, labeled by blah.”
That’d be great UX.
You are falling into a common trap. LLMs do not have understanding - asking it to do things like convert dates and put them on a number line may yield correct results sometimes, but since the LLM does not understand what it’s doing, it may “hallucinate” dates that look correct, but don’t actually align with the source.
Thank you for calling that out. I’m well aware, but appreciate your cautioning.
I’ve seen hallucinations from LLMs at home and at work (where I’ve literally had them transcribe dates like this). They’re still absolutely worth it for their ability to handle unstructured data and the speed of iteration you get – whether they “understand” the task or not.
I know to check my (its) work when it matters, and I can add guard rails and selectively make parts of the process more robust later if need be.
Arc has an LLM that lets you replace your search functionality with search or ask, where if you type a question it tries to answer it based on the content on the page. Kinda close to what you’re talking about.
Arc is genuinely trying to use LLMs in their browser in interesting ways.
Webpage authors use LLMs to generate extremely long articles, to make you scroll by ads for longer. You use LLMs in your browser to summarize those articles. The circle of life, or something.
I stumbled upon a website through DDG, and after a long intro, the main section supposedly where the thing I was searching for had “Sorry I can’t fulfill your request right now”. Basically a fully generated page to match my search with some parasitic seo tactics. The web be chaging. Front page of DDG.
Edge is branding itself “The AI Browser”. Chrome has plans to embed LLMs for text input. Opera, the browser which was commandeered from the original Vivaldi team and turned into a crypto/VPN gimmick browser, is of course among the hardest leaning into the LLM trend.
Quick tool to summarize a page, proofread, or compare it to another source. Still needs a functioning human brain to separate the wheat from the chaff so to speak, but I could see a LLM (especially local) being useful in some ways.
I’m sure there are disabilities or unique use cases that could increase it’s usefulness, especially once they improve more.
