[paxys]

So you click a button, it pops open a text box in a floating window, you type in a question, and the AI replies. This is the most underwhelming implementation of browser-based AI that they could have come up with. Quite literally just gemini.google.com in an iFrame.

[qnleigh]

> Quite literally just gemini.google.com in an iFrame.

Hmm, no? It has access to all of the content of all of you're currently open tabs, and is able to parse images on web pages as well.

It would be neat if it could also browse on your behalf, but that would present all kinds of security risks.

[paxys]

No, it can only access the tab you are currently on. And that too just the content that is already available. It can't scroll up and down to load more. It can't follow links. It can't run any actions. You'll get a ton more functionality by just taking a screenshot of the page yourself and pasting it in ChatGPT/Claude/Gemini.

[IX-103]

I'm sure that kind of functionality is coming. There's a lot of activity in the chromium repo (chrome/browser/actor/tools) that appears to be adding support for that sort of orchestration.

[rhetocj23]

Ok cool.

But whats the vision of this? Where are they trying to take the customer?

I feel like this issue relates back to the origin of Google (search) in the first place. It was borne out of a technology in which the founders did not envision what it would become. It seems the firm just tries ideas and then tries to figure out where it goes - thats the culture. And unsurprisingly, yields a lot of failiures.

In contrast, Apples approach yields a much higher rate of success with less risk.

[troyvit]

I feel the same about Firefox's vision, although I admit I haven't tried it. Often when I visit a place like chat.mistral.ai Firefox gives a weird popup that says something about "don't you wish you didn't have to open this in a tab?" Like is that their AI vision? Saving me a tab?

[rhetocj23]

Ive come to realise the constraint on innovation et al is not external resources - its what comes from within. Imagination.

[thwarted]

We're too lazy to browse now, that we need machines to do it?

[dheera]

When every goddamn webpage presents itself to me like this: https://i.imgur.com/mMi8an3.png

I do need machines to do the browsing for me.

[x______________]

You might want to explore ad, script and popup blockers(EG: no script, ad block plus, ghostery).

They exist for that very reason, the web is much friendlier.

[troyvit]

No no no we don't need a sustainable answer to the cancer of ads on the internet, that would break capitalism and send the world sliding into chaos! No, see, what we need is AI in our browsers. That is going to transform things.

[mock-possum]

I don’t think that website actually wants you to be here, it looks like it’s trying to get you to leave.

[typpilol]

You need an ad blocker badly lol

[riffraff]

The idea is you could ask a to browser to do things like operate on multiple websites to do boring stuff, e.g. cross check phone reviews across sites x y and z.

I 100% don't feel comfortable letting my browser work alone, but "agentic browsers" are a thing some people want and/or are building.

[baq]

A small part of me wants this to spectacularly succeed so I can stop using whatever the army of figma designers wishes to force down my throat when most things I need could be spreadsheets with a few buttons with macros hooked up.

[ehnto]

It makes sense as an avenue for Agents as well, since it is the defacto "work app" platform. For many, their entire workday is spent inside the browser.

[garyfirestorm]

It’s not browsing we are lazy at. It’s parsing through ton of results until we find what we were looking for.

[blharr]

>It has access to all of the content of all of your currently open tabs

This is supposed to be a good feature? Not a privacy nightmare?

[iansinnott]

Likely depends on whether or not its opt-in. If Gemini only gets page content when you ask it to then that's fine.

Of course, it should also be possible to completely disable Gemini so as to avoid accidentally sending it private browsing content.

[weatherlite]

> So you click a button, it pops open a text box in a floating window, you type in a question, and the AI replies. This is the most underwhelming implementation of browser-based AI that they could have come up with. Quite literally just gemini.google.com in an iFrame.

Well, they're gonna have to support an astronomical scale of queries - not many companies in the world are able to do it and Alphabet is doing it pretty much on their own stack of cloud, a.i chips and software. So sure, the front end is not a big deal but this is still a big move.

[atdt]

It has access to the current page, so you can ask Gemini questions about its content.

[milkshakes]

don't worry -- they have big plans!

https://research.google/blog/mechanism-design-for-large-lang...

[j_timberlake]

They took 1 step at a time instead of trying to take multiple steps at a time, how is that a bad thing. They're obviously getting things prep'd for Chrome agents and Gemini 3.

[nomilk]

I think it will be useful as a modernised ctrl+f