Top
Best
New

Posted by AndrewDucker 4 days ago

ChatGPT's Atlas: The Browser That's Anti-Web(www.anildash.com)
756 points | 327 commentspage 5
dangoodmanUT 1 day ago|
I think the idea of "we're returning to the command line" is astute tbh, I've felt that subconciously and I think the author put it into words for me.

The article does taste a bit "conspiracy theory" for me though

chartered_stack 1 day ago||
I think we're returning to CLIs mostly because typing remains one of the fastest ways we can communicate with our computers. The traditional limitation was that CLIs required users to know exactly what they wanted the computer to do. This meant learning all commands, flags etc.

GUIs emerged to make things easier for users to tell their computers what to do. You could just look at the screen and know that File > Save would save the file instead of remembering :w or :wq. They minimized friction and were polished to no end by companies like MSFT and AAPL.

Now that technology has got to a point where our computers now can bridge the gap between what we said and what we meant reasonably well, we can go back to CLIs. We keep the speed and expressiveness of typing but without the old rigidity. I honestly can't wait for the future where we evolve interfaces to things we previously only dreamt of before.

ewoodrich 1 day ago||
It’s less rigid than a command line but much less predictable than either a CLI or a GUI, with the slightest variation in phrasing sometimes producing very different results even on the same model.

Particularly when you throw in agentic capabilities where it can feel like a roll of the dice if the LLM decides to use a special purpose tool or just wings it and spits out its probabilistic best guess.

chartered_stack 22 hours ago||
True the unpredictability sucks right now. We're in a transition stage where the models can understand intent but cannot constrain the output within some executable space reliably.

The bridge would come from layering natural languages interfaces on top of deterministic backends that actually do the tool calling. We already have models fine-tuned to generate JSON schemas. MCP is a good example of this kind of stuff. It discovers tools and how to use them.

Of course, the real bottle neck would be running a model capable of this locally. I can't run any of models actually capable of this on a typical machine. Till then, we're effectively digital serfs.

dangoodmanUT 1 day ago|||
that being said, asking chatgpt to do research in 30 seconds for me that might require me to set aside an hour or two is causing me to make decisions about where to tinker and ideas to chase down much faster

can never go back

h4ck_th3_pl4n3t 2 hours ago||
That's great, you're already addicted. The perfect regular user.
pimlottc 1 day ago||
It’s not so much a conspiracy theory as it is a perfect alignment of market forces. Which is to say, you don’t need a cackling evil mastermind to get conspiracy-like outcomes, just the proper set of deleterious incentives.
hermitcrab 14 hours ago||
'Atlas' as in 'Atlas Shrugged' by any chance? If so, ugh.
PaulHoule 15 hours ago||
One could make the case that the web of 2025 is anti-human. AI clients are one of very few exits we have from enshittification. ChatGPT can read all the ads so you don’t have to. The whole point of those annoying CAPTCHAs and those stupidity anime girls on all the kernel web sites is slamming the door from any exit from Google’s world that we live in.
falcor84 23 hours ago||
It really lost me at

>There were a tiny handful of incredible nerds who thought this was fun, mostly because 3D graphics and the physical touch of another human being hadn't been invented yet.

I can barely stomach it with John Oliver does it, but reading this sort of snark without hearing a British voice is too much for me.

kleiba 22 hours ago|
Also, re: "a tiny handful of incredible nerds" - page 20 of this [0] document lists the sales figures for Infocom titles from 1981 to 1986: it sums up to over 2 million shipped units.

Granted, that number does not equal the number "nerds" who played the games because the same player will probably have bought multiple titles if they enjoyed interactive fiction.

However, also keep in mind that some of the games in that table were only available after 1981, i.e., at a later point during the 1981-1986 time frame. Also, the 80s were a prime decade for pirating games, so more people will have played Infocom titles than the sales figures suggest - the document itself mentions this because they sold hint books for some titles separately.

[0] https://ia601302.us.archive.org/1/items/InfocomCabinetMiscSa...

brendoelfrendo 1 day ago||
Atlas confuses me. Firefox already puts Claude or ChatGPT in my sidebar and has integrations so I can have it analyze or summarize content or help me with something on the page. Atlas looks like yet another Chromium fork that should have been a browser extension, not a revolutionary product that will secure OpenAI's market dominance.
coffeefirst 22 hours ago||
Yep. I was playing around with both Atlas and Comet and, security and privacy issues aside, I can’t figure out what they’re for or what the point is.

Except one: it gives them the default search engine and doesn’t let you change it.

I asked Atlas about this and it told me that’s true, the AI features are just a hook, this is about lock in.

Make of that what you will.

h4ck_th3_pl4n3t 2 hours ago||
It is not a fork, and it is a NewTab page Chrome extension.

They didn't even change the codename that's displayed in the Settings page.

thetrees 10 hours ago||
Condensed version: 'OpenAI built a browser that mostly benefits OpenAI. Instead, they should have built a browser that mostly benefits the customer (and, thus, benefits OpenAI less).'

Welcome to capitalism.

I wish it wasn't like this, but this is the system we have. If you want OpenAI to do things differently, another company has to do it differently, show that doing it differently is even more profitable (see Apple and their privacy angle), and force OpenAI to change for competitive reasons.

Or figure out some way to trick our way into a constitutional convention so we can change the system.

roundatlas 1 day ago||
What now remains is, after hearing glowing feedback, Satya making this the default browser in Windows as part of Microsoft and OpenAI's next chapter.
hereme888 22 hours ago||
> The fake web page had no information newer than two or three weeks old.

What irks me the most about LLMs is when they lie about having followed your instructions to browse a site. And they keep lying, over and over again. For whatever reason, the ONE model that consistently does this is Gemini.

kingkawn 18 hours ago||
People didn’t hate DOS it just was what it was
dtagames 1 day ago|
This article is deep, important, and easily misinterpreted. The TL;DR is that a plausible business model for AI companies is centered around surveillance advertising and content gating like Google or Meta, but in a much more insidious and invasive form.

Worth reading to the end.

brabel 1 day ago|
I found the article is no more than ranting about something that they are just projecting. The browser may not be for everyone, but I think there’s a lot of value to an AI tool that helps you find what you’re looking for without shoving as many ads as possible down your throat while summarizing content to your needs. Supposing OpenAI is not the monster that is trying to kill the web and lock you up , can’t you see how that may be a useful tool?
More comments...