Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

ChatGPT Atlas

chatgpt.com

Post reply on HN
Print view View original

351–360 of 768 posts

Re: ChatGPT Atlas

#351

Post by dotancohen »

EGreg wrote: post #219
hbn wrote: post #192

Earlier quoted context omitted.

After skimming the product page I'm still not sure what extra data exactly everyone is so confident is being gathered/used in a way that Google wouldn't already be doing in Chrome. As far as I can tell, most of these features are already integrated into Chrome but with Gemini. What exact feature in Atlas would need to log your every keystroke? Could they be doing that? Yes. But so could Google and in both cases they'…

Our trusted computing base should be small, built from open-source, and not under the control of one company. But sadly, here we are. How do we know GMail can't steal your bank account info and Chrome can't steal ... everything from your web browsing, or impersonate you? All they have to do is be pressured by a government: https://www.aph.gov.au/Parliamentary_Business/Bills_Legislat...

Could you mention where in that bill there is concern for the Australian government pressuring tech companies for the ability to impersonate you? Thank you!

Re: ChatGPT Atlas

#352

Post by gulfofamerica »

tempestn wrote: post #323
sloankev wrote: post #306

Im still weary of OpenAI being legally required to retain all of your data even if you delete it [0] . This means everything you expose to this tool will be permanently stored somewhere. Why isn’t this a bigger problem for people? Even privacy concerns aside… this would be the world’s most catastrophic data leak. [0]: https://openai.com/index/response-to-nyt-data-demands/

I think you mean wary, not weary.

Por que no los dos.

Re: ChatGPT Atlas

#353

Post by Ethee »

elric wrote: post #350
granzymes wrote: post #279

Earlier quoted context omitted.

Being able to search browser history with natural language is the feature I am most excited for. I can't count the number of times I've spent >10 minutes looking for a link from 5 months ago that I can describe the content of but can't remember the title.

Are we talking searching the URLs and titles? Or the full body of the page? The latter would require tracking a fuckton of data, including a whole lot of potentially sensitive data.

All of these LLMs already have the ability to go fetch content themselves, I'd imagine they'd just skim your URLs then do it's own token-efficient fetching. When I use research mode with Claude it crawls over 600 web pages sometimes so imagine they've figured out a way to skim down a lot of the actual content on pages for token context.

Re: ChatGPT Atlas

#354

Post by echelon »

motoxpro wrote: post #236
echelon wrote: post #42

Earlier quoted context omitted.

If the DeepSeek approach to training hyperscaler models "cheaply" after all the hype wears down works, we just need to follow in their footsteps and build open source alternatives to everything. Frontier models take a lot of money and experimentation. But then people figure out how to train them and knowledge of those models and approaches leaks. Furthermore, we can make informed guesses. But best of all, we can exfi…

Curious as to what companies you think DO have a moat? I would say their moat is the same as Facebook's. Not permanent (see TikTok) but also super strong.

TikTok, Facebook, Instagram, etc. have network effect stickiness. YouTube has a vast library that grows in value over time and has been able to lose money for long enough to make the experience unassailable, etc. (Look at Vimeo.)

Maybe LLMs get that by knowing our entire past? But I find that creepier than useful. Right now ChatGPT is at the top of the world, but I don't see it becoming the new unrivaled Google Search. There's just too many people building it, and once OpenAI starts monetizing and "enshittifying" it, other offerings will become more compelling.

AI models put a large swath of mostly tech companies at risk. Including the old business models of titan products like Google Search.

Image/video/world models specifically do this more to legacy media incumbents than LLMs do to complex business processes. We see orders of magnitude savings with marketing, design, film, and possibly in the future game design. LLMs, on the other hand, aren't great at getting your taxes or complex business logic right.

Re: ChatGPT Atlas

#355

Post by baal80spam »

jacekm wrote: post #331
granzymes wrote: post #279

Earlier quoted context omitted.

Being able to search browser history with natural language is the feature I am most excited for. I can't count the number of times I've spent >10 minutes looking for a link from 5 months ago that I can describe the content of but can't remember the title.

I think that such feature is already available in Chrome https://support.google.com/chrome/answer/15305774?hl=en

Ah, makes sense why I need to use an extension for that:

> To use this feature, you must be located in the US

Re: ChatGPT Atlas

#356

Post by andoando »

seizethecheese wrote: post #197
gh0stcat wrote: post #91

The thing I find the most funny about all of these demos is they outsource tasks that are pretty meaningful ... choosing where to hike, learning more about the world around you, instead, you'll be told what to do and live in blissful ignorance. The challenge of living life is also the joy, at least to me. Plus I would never trust a company like openai with all of my personal information. This is definitely just them…

I find the hike itself more meaningful than the searching for it. If an LLM can recommend be me a better hike, I’m all for it. The word choice here: “you’ll be told what to do” doesn’t really reflect my experience with LLMs. You can always ask for more recommendations or push back. (As an aside, I’ve found LLMs to be terrible for recommending books.)

I found a big part of what makes doing an activity enjoyable is the time spent thinking/planning involed to make it happen.

For example if I spent a week looking at exactly how to plan my trip, and then finally going out to accomplish it vs just waking up one morning and someone guiding me on exactly what to do

Re: ChatGPT Atlas

#357

Post by jmkni »

sansseriff wrote: post #20

I haven't used LLM chrome plugins because I couldn't trust that they weren't collecting more information about my browsing than I'd like. The same concern exists for this, though now I'm just confident it's a giant software company with access to my data rather than some shady plugin developer. I'm faced with asking myself if that's actually better...

Yeah I find LLM's very powerful in the right context, but I like to keep them at arms length

I will go to them when I need something, instead of them spying on me incase I need something

Re: ChatGPT Atlas

#358

Post by granzymes »

sloankev wrote: post #306

Im still weary of OpenAI being legally required to retain all of your data even if you delete it [0] . This means everything you expose to this tool will be permanently stored somewhere. Why isn’t this a bigger problem for people? Even privacy concerns aside… this would be the world’s most catastrophic data leak. [0]: https://openai.com/index/response-to-nyt-data-demands/

Thankfully the New York Times lost their attempt to force OpenAI to continue preserving all logs on an ongoing basis, but they still need to keep some of the records they retained before September.

https://mashable.com/article/openai-court-ordered-chat-gpt-p...

Re: ChatGPT Atlas

#359

Post by mk89 »

jryio wrote: post #15

If you think this is useful... remember technology like this would make totalitarian leaders foam at the mouth. If you thought that ads are creepy, Atlas is a root level keylogger service. Why would you want an AI company scraping and recording all of your browser interactions. Yes Google already does this via Chrome. It's one thing to build a predictive model on your demographic, spending, location and income inform…

Man, literally everything we have been doing since the 90's makes totalitarian leaders foam at the mouth.

Do you understand we're willingly sharing our name, surname, relationships, friends, where we work, what we do, how much we make (not maybe precisely, but with some social engineering you can get that), in some cases even private intimate videos, pics of our families, etc. Everything.

There is nothing else they need anymore. If they want to, they get you. Any time. And yet, things work relatively well.

Re: ChatGPT Atlas

#360

Post by elric »

It's very ironic that we seem to be getting "agentic browsers" before we got browsers that are actual "user agents". Browsers have become slaves to the websites they render, ostensible in order to protect users. But they don't let you do simple things like decide which form fields to remember (like some sites telling the browser not to remember the input of the username field), they prevent users from copy/pasting text. Etc. But throw in some LLM foo and suddenly we're golden? Pah.
Post reply on HN