'ChatGPT's Atlas: The Browser That's Anti-Web' (anildash.com)
- Reference: 0179906868
- News link: https://it.slashdot.org/story/25/10/29/177230/chatgpts-atlas-the-browser-thats-anti-web
- Source link: https://www.anildash.com/2025/10/22/atlas-anti-web-browser
> When I first got [2]Atlas up and running, I tried giving it the easiest and most obvious tasks I could possibly give it. I looked up "Taylor Swift showgirl" to see if it would give me links to videos or playlists to watch or listen to the most popular music on the charts right now; this has to be just about the easiest possible prompt.
>
> The results that came back looked like a web page, but they weren't. Instead, what I got was something closer to a last-minute book report written by a kid who had mostly plagiarized Wikipedia. The response mentioned some basic biographical information and had a few photos. Now we know that AI tools are prone to this kind of confabulation, but this is new, because it felt like I was in a web browser, typing into a search box on the Internet. And here's what was most notable: there was no link to her website.
>
> I had typed "Taylor Swift" in a browser, and the response had literally zero links to Taylor Swift's actual website. If you stayed within what Atlas generated, you would have no way of knowing that Taylor Swift has a website at all.
>
> Unless you were an expert, you would almost certainly think I had typed in a search box and gotten back a web page with search results. But in reality, I had typed in a prompt box and gotten back a synthesized response that superficially resembles a web page, and it uses some web technologies to display its output. Instead of a list of links to websites that had information about the topic, it had bullet points describing things it thought I should know. There were a few footnotes buried within some of those response, but the clear intent was that I was meant to stay within the AI-generated results, trapped in that walled garden.
>
> During its first run, there's a brief warning buried amidst all the other messages that says, "ChatGPT may give you inaccurate information", but nobody is going to think that means "sometimes this tool completely fabricates content, gives me a box that looks like a search box, and shows me the fabricated content in a display that looks like a web page when I type in the fake search box."
>
> And it's not like the generated response is even that satisfying.
[1] https://www.anildash.com/2025/10/22/atlas-anti-web-browser/
[2] https://tech.slashdot.org/story/25/10/21/1725235/openai-debuts-ai-powered-browser-with-memory-and-agent-features
Who cares about this fanboi's take? (Score:1)
What if it told him she's shallow and untalented? Would he have had a heart attack?
Re: (Score:3)
I mean the combination of Taylor Swift and AI is the perfect storm of contemporary pointlessness... It's hairdressers and telephone sanitisers all over again.
Re: (Score:2)
I think it was intended as a supremely milquetoast query that would have a search engine pretty much pop up a specific thing the user is after.
And the LLM first approach is *really* bad at that. If you are looking for an existing, canned piece of content, the LLM is a let down. A large chunk of what people want is an existing thing.
LLM as a readily available *option* for the sorts of inputs that it works with? Sure. As a replacement for internet search, not so much.
Re: Who cares about this fanboi's take? (Score:2)
How much do you want to bet that Taylor Swift fans won't care because they're happy with consuming whatever crap their corporate overlords shovel at them?
It's just the Facebook model (Score:5, Insightful)
Instead of linking you to the information at the source, Facebook (and Twitter, and ...) tries does its best to keep you inside its ecosystem. That way you still see their advertisements, they can still collect your data for monetization, etc.
It's the approach most of the major players are trying to emulate now.
Re: (Score:2)
I can't think of another technology that can be so wrong and yet gets pushed at everyone.
Re: (Score:3)
Yibble, the nearest flumble I can thunk up is the nutri-goop splinklefluff industry — all shimmer, no sustenance, and plenty of jazzy jars of who-knows-what!
Re: (Score:3)
> If you want information about Taylor Swift why wouldn't you just use a regular search engine?
Evidently, that's what was done. But Atlas takes you somewhere else:
> I had typed "Taylor Swift" in a browser, and the response had literally zero links to Taylor Swift's actual website.
As others have stated, AI is a move back toward a closed ecosystem. If Taylor Swift wants to be found, she can just buy an ad with each LLM service.
When one uses a search "AI assistant" for a "summary" the first thing I notice is that all references, citations, links to other web sites are stripped off. AI will be the new Ministry of Truth. No need to look any further, citizen.
Re: (Score:2)
We've always been at war with Eurasia.
Oh no! (Score:2)
AI once again proves itself to be over-hyped shite.
Anyway...
Deja Vu Hallucinations All Over Again (Score:2)
Remember when all we could talk about was AI slop? Then they started fixing that stuff one-by-one. Well, next up is browsers.... and they'll fix that stuff one-by-one. *yawn*...
Just another shitty Chromium fork? (Score:1)
Fuck off Altman, you sleazy crypto scammer.
Oh no you didn't! (Score:5, Interesting)
Anil Dash dares to dip on Zork???
"There were a tiny handful of incredible nerds who thought [Zork] was fun, mostly because 3D graphics and the physical touch of another human being hadn't been invented yet. "
How dare you sir! HOW DARE U
It's clear that Anil never found the mailbox or the house or the window that's slightly ajar and just stood there in the clearing arguing with the CLI until rage quitting the game.
"We left command-line interfaces behind 40 years ago for a reason"
You left, Anil, you left. Clearly the experience with Zork was a traumatic event in your life. The CLI gave you bad feelings and you gave up on learning.
The important bit in the article is this one:
"During setup, Atlas pushes very aggressively for you to turn on "memories" (where it tracks and stores everything you do and uses it to train an AI model about you) and to enable "Ask ChatGPT" on any website, where it's following along with you as you browse the web. By keeping the ChatGPT sidebar open while you browse, and giving it permission to look over your shoulder, OpenAI can suddenly access all kinds of things on the internet that they could never get to on their own."
That's extremely bad. For the benefit of some kind of convenience or productivity boost, Atlas wants to remember everything you've done, so you don't have to.
I know some people will say how incredible this is as a feature, but you'd be wrong. Nobody needs that. 99.9% of the time that's not a useful thing to have, considering what it costs in privacy and just overall fuck OpenAI, Sam Altman and his kind who want to watch the human race devolve as they feed it more garbage, all the while training some "super genius" AI that will tell them the secrets of the universe and how to be immortal trans-humans. But obviously just tell them the secrets, not the rest of us, because we're going to have to pay for the privilege they've foisted on us as a necessity now that the planet is on fire due to the CO2 these giant data centers are producing.
Sam Altman believes he can become a god, and AGI is the way to achieve that.
The Ass-Retard Quote of the Day (Score:1)
"We left command-line interfaces behind 40 years ago for a reason"
How did that work out for us? Take a look around. The whole world is a slow motion train wreck because of AI and iPhones.
Give me a pound of TARDmeat. Sliced thick.
Indeed (Score:2)
Nothing about "Prince Harry asked about Taylor Swift and Charli XCX feud"
No "What did Travis Kelce say about proposing to Taylor Swift?"
Nothing about the 12+ boyfriends she had?
Indeed useless.
12 year olds won't use Atlas then I guess.
Well ... (Score:2)
... in one sense, this was part of the promise of the web - that "user agents" of all sorts would be able to intake the structured information from HTML pages, and present it in whatever way was useful to the user.
I'm not saying this particular thing is useful ... just that the underlying concept isn't so alien as some seem to think.
I have not found a use case for... (Score:2)
...an AI browser
Maybe someday, as they evolve, I might find them useful, but today, I see no value
I am surprised by the amount of hype around them
This thing is comically bad (Score:2)
Sure, let the tripping robot have access to your bank account while watching it use a browser like an 80 year old who has never seen a mouse.
I'll watch.
Re: (Score:1)
"It doesn't matter what you think - we've had a fantastic amount of investment - partly based on artificially-inflated value created by roundtripping - we are going to succeed - over your dead body if needed" someone might have said.
Re: This thing is comically bad (Score:2)
Remember when online banking was unthinkably insecure but banks just paid for mistakes with profits from equity investments which increased because people were using online banking?