Putting Gemini to Work in Chrome
blog.google
blog.google
Are there really people who are like "Man, if only I could this straight in Chrome" ? Is this something worth bloating a browser (further) with?
You know, the whole "faster horse" argument.
Of course the single biggest thing my browser can do to help me is blocking ads, which means it's curious to see this just after Google killed adblock in chrome.
Even before the removal of MV2, the claims that it would kill adblock were ridiculous as many adblockers had already switched to MV3 but it was at least understandable that people could be ignorant of that fact. Now that everything is on MV3 how can people still be claiming that Google killed adblock when Chrome users still have working adblockers?
However, Gemini in Chrome requires you to allow them to use your data to improve their model, which I won't consent to. Google workspace account seems exempt so I plan to try it out there.
These feels on par with Microsoft's push to shove Copilot down everyone's neck at every step possible whether we like/need it or not
Also for things like locating a product available to pick up in a nearby store, it’s crazy how often Google fails at this particular task.
Anyways I will (continue to) not touch Chrome with a 1000ft pole after this. AI is awful in almost every aspect I've ever tasked it to.
You can always ask for a list and still make the final decision yourself.
Sounds like a contrived situation, but there's a surprising amount of "thought leader" CEOs out there who make completely nonsensical decisions under the banner of "saving costs and automating things".
(Real-world example I know of) company pays for cheapest tier they can find of Gemini, tell everyone to use it. But won't pay for Asana seats, so every user in your 100-person startup is a guest, and can't use the connector in any AI app to TRULY do useful task management with AI.
Having some better access to AI in the browser would pave over that pain for someone who currently doesn't want to spend their own money on something like Claude for Cowork and the Chrome extension to drive the browser, or open a terminal to have Claude Code do it.
I dread having to log in to these systems and waste hours achieving the simplest tasks.
This is what I'm using Claude for. E.g. I log in to AppStore connect, tell it what I need (3 subscription tiers), it will do all the clicking and editing and Apple's stupid UI, then I will ask it to create a summary for RevenueCat, and use another Claude session in there to click all the buttons to configure based on what just happened in Appstore connect.
Or configuring S3 buckets or whatnot.
Examples: using my budgeting app directly to figure out why some forecasting event went wrong, or helping me correlate SOC2 tickets with GitHub pull requests and flagging all that are older than $date.
It’s surprisingly convenient for a narrow set of tasks.
I know it calls out that there’ll need to be user confirmation before the final purchase, but if you’re already not expending the effort to find the product or service yourself, are you really going to sit and research what it’s given you? If you are, then what’s the point of using the agent?
Just seems like the next evolution in Google’s ad revenue generation.
I think the vision here is for your browser agent to spend money with google's partners on pointless consumer slop while you sleep.
What I just discovered: The good old google search without AI bloat but with privacy via https://www.startpage.com. Highly recommended!
I'm having a hard time understanding why I will tell gemini to create an account on some website for me or send an email. Those are usually just a tab away. That's why I feel like I'm missing something here.
But for regular browsing? I don't see the point.
Thats exactly what an agent regularly fails at.
Like the first example in the demo carousel (the Y2K party) starts from a photo and a prompt of roughly "buy the props needed for replicating this photo from Etsy". It first analyzes the image in the current tab, identifies a bunch of things to buy, searches for them on Etsy, customizes the orders, adds them to the shopping basket, and then asks for a confirmation to actually send an order.
The second one auto-fills a form with a couple of dozen fields from the data that's in a pdf in another tab. (And in the fiction of a demo, presumably a pdf that's you already had around, not one that you made just for the purposes of using it to auto-fill the form.)
I'm not the target market for this: automating a browser with my credentials is just too scary, but I can certainly see the utility. There's a huge amount of tasks taking a minute or two are not worth creating bespoke automation for but that are also pretty mechanical processes.
It’s as if they used AI to generate use cases for their AI tool because they weren’t really sure what it’s for…
I suppose by being in the browser it can private and paywalled data, so maybe that's something.