[reisse]

I see the vision here, which the top commenters (sorry, couldn't read all of them) seems to miss. This should be a moonshot bet on the next generation of user experience. People are complaining about apps, but the idea here should be to make apps irrelevant as a concept. You don't need "apps", you need data feedable to LLM and a visualization toolkit for presenting results. And maybe some tools to manually wrangle the data when precise manipulation is required.

On paper, this sounds amazing. Like "out of sci-fi books" amazing. The caveat, though? I very much doubt Google has the capacity to execute this properly. And we'll get another half-baked attempt at reskinning Chromium and/or Android.

[isityettime]

I can't get Gemini to not cut me off mid-sentence and reply to only half of what I wanted to say.

It doesn't reliably integrate with Google Maps.

I can't trust it for basic shopping tasks because it doesn't reliably evaluate stated criteria. I can't even ask it if a specific web store has a specific item in stock and trust the answer.

Google hasn't gotten the bare-ass basics right. They're not about to "make apps irrelevant as a concept".

> [W]e'll get another half-baked attempt at reskinning Chromium and/or Android.

100%!

[joemazerino]

This product is exactly how Google improves Gemini. For better or worse (I'd argue worse) Google will be putting these devices in classrooms and everywhere else to train Gemini.

[nobrains]

My take is related but different. I believe that Google believes the customer's for these "premium" Googlebooks are the kids who used Chromebooks all their school years, and are now graduating school, getting into university, or getting into their first job. Google is presenting them an option, that feels familiar, but is also premium. The next level. The big-boy upgrade. Familiar but Better.

[rchaud]

Except that Chromebooks are widely reviled as cheap, laggy garbage plus a negative mental association with tedious schoolwork. If anyone buys one, it'll be in spite of the brand association.

[ajross]

> I can't get Gemini to not cut me off mid-sentence and reply to only half of what I wanted to say.

This kind of criticism is unpersuasive: "USENET and ftp are fine! Why do I need this monster 'mosaic' thing which needs 8MB of core just to try to see some buggy 'dynamic page' someone threw together where the server is always down anyway!"

New technology is incomplete, rushed to market, inconsistently reliable and rapidly improving. News at 11!

That doesn't mean Googlebook is going to take over the world. But no, you don't get to live in 2019 forever either.

[gumby271]

I would say we're reaching new levels of incomplete and rushed to market though. The original Google Search, Maps, Docs, etc. weren't feature rich, but what they had worked well. I'm not saying they were perfectly built, but they weren't falling apart to the degree Gemini and other Google products are. And Google gets to do this now, no one can compete or leave them, so why do anything well.

When Google first added Gemini to Sheets, I tried it and it completely hallucinated transaction data on the bank statement I was testing it on. It's worse than incomplete, it's broken and dangerous. But Google gets to slap a disclaimer on Gemini that says "results may be fucked" and they get a pass to ship it.

I have no expectation that Google can improve its new products since everything they've been doing is going in the opposite direction. You're not seeing pushback on embracing new technology, its purely about Google not being trusted to do things well.

[ajross]

> I would say we're reaching new levels of incomplete and rushed to market though

The year 2001 called, it was laughing hysterically and I couldn't quite make out what it was saying.

[gumby271]

I think I'm too stupid to understand what point you're trying to make, sorry.

[ajross]

It was that the AI boom/bubble is a new technology in its infancy. It's not comparable to the stuff you mention, which were all[1] second wave "Web 2.0" products being introduced into an established market. The proper comparison would be to the period in the mid-late 90's up until the dot com burst, when "internet" technology was sudden and gestational.

And it sucked. Nothing worked. TCP was "established" technology that needed tweak after tweak. The whole stack failed regularly and catastrophically, at every level from backbone routers to browser rendering. People were gluing together solutions to problems using tools written on 1970's teletypes because the "new" stuff from Microsoft was even worse.

It was glorious. But it sucked. And the same is true right now. Nothing you use is going to do what you want exactly. And we all get to figure out how to make it not suck together.

[1] Except Google search itself, which was a comparative latecomer (anyone remember DEC and Altavista?) but still part of the initial rush. And thus is sort of the exception that proves the rule: the one product in the infancy rush of a new technology wave that doesn't suck (or sucks least) gets to be the next tech giant. Seems like the smart money right now is on Anthropic, but we'll see.

[jaapz]

Youre looking at the past with rose tinted glasses.

[isityettime]

I never mentioned any alternative to the Googlebooks, let alone said something like "$PRECEDING_PRODUCT is fine". You're refuting a kind of categorical naysaying that I haven't actually done.

Maybe some day, someone will make a computer with integrated digital assistants that virtually everyone really wants. Maybe that someone will even be Google.

But I wasn't writing sci-fi or putting together a pitch deck on Google's behalf. I was considering the actual technology (and priorities) they have on offer.

[mrhottakes]

Your criticism is unpersuasive; you haven't laid out any actual argument, just sarcasm.

[disqard]

Your argument has merit.

I'm just going to point out that NCSA Mosaic was released in 1993, and the hand-wringing over it being an absolute memory pig is more 1999 than 2019.

But your overall point is well-taken.

[ajross]

FWIW, I wrote 2019 as a representation of "last year before AI mattered" and not when people cared about the early web. Also by 1999 Mosaic had been forgotten and everyone was running Netscape and IE4. :)

To wit: complaining about AI software quality right now is isomorphic to whining about how much the early web sucked. It did suck. But it was still the future.

[rkozik1989]

You forgot the part where average consumers kind of hate AI and won't buy this or use that feature because of its AI.

[ulimn]

I'm not sure about that. The "normies" around me love AI. My mother and mother-in-law drive me crazy when I ask for advice and they copy the answer directly from ChatGPT. Not to mention the stupid images generated by AI.

[MattRix]

In my experience, old normies love AI, young normies do not.

[mywittyname]

My parents hate AI, as do my nieces and nephews.

It's not an age thing. It seems to be an experience thing. If AI is perceived to have ruined something you like, or otherwise negatively impacted your life, you're going to hate it.

[10729287]

The conclusion here is that we’re all someone else’s normie I guess.

[LastTrain]

I’m not sure what you mean by “normies” - non technical people? I don’t know a single person who /loves/ AI. I know some that speculate about potential benefits. I know some that use it begrudgingly. Some that have some anecdotes of it being useful but mostly couched in dread. Almost universal is the feeling that it is being forced on them in ways they do not want. Who loves it?

[tim333]

There's stuff I think if cool but is not really regarded as AI any more like language translation and I find the google lens thing in Chrome very handy - does ocr, finds stuff in images etc. Maybe people will like more the stuff that doesn't get called AI but just does useful things?

[remus]

I think it is hard to generalise like this. There's a lot of people in the world and we each only know a very small slice of them. I happen to know several people who love AI, I don't feel like have a good grasp of the where the overall sentiment is.

[mywittyname]

> I'm not sure about that. The "normies" around me love AI

Really? Because every "normie" I know complains to me about how much they hate AI.

Lots of the the complaints center around it being forced on users, how all information on the web is AI generated anymore, thus can't be trusted, broader issues around how it makes society much less stable (bots, misinformation, fakeporn, and more recently, breaking security). And, of course, those whose jobs are actively being replaced by AI really hate it.

The data centers going up locally are likely to be another avenue for complaints.

[u_fucking_dork]

The average Redditor that performatively hates AI sure. But outside of the chronically online bubble, the average person, in my experience, does not hate AI

[NoGravitas]

Surveys show otherwise. As does the viral video of that lady giving the AI-booster commencement speech and getting booed.

[dymk]

How confident are you that those surveys portray the “average” person? ChatGPT was the fastest adopted app… ever.

[jorl17]

My average tech friend kind of hates AI.

My average "non-tech" friend or acquaintance? Absolutely loves it and is using it daily.

So, not my experience AT ALL.

[fortyseven]

I think once the hype dies down, and things start to normalize, the hate will fade into the background. Instead of "AI", we'll just get "apps with features" again.

[spandrew]

This is not correct. People trust their "apps". They know Spotify's algo can feed them novel music. They know Kobo's reader connects to libraries in a way that isn't the DRM-lockin that Amazon shovels.

Most non-tech folks are incredibly skeptical about AI due to the piss-poor brand management of OpenAI and Anthropic—pushing an AGI Skynet post-employment narrative to make the tech feel bigger than it is.

I trust Google (and certainly Apple) to figure some of this stuff out—but nobody wants an LLM, that nobody knows how it operates, to run the show entirely. Heuristically, we may get away with a Human-on-the-loop approach—-not a Human-in-the-loop or Out-of-the-loop approach.

[cloverich]

IMO the right next step, which Apple will get to slowly, is better Siri and better app intents. No apps reminds me a bit of the no folders, use search for everything and organize nothing, approach. To me it falls flat because structure - which apps provide - acts as context, memory, mental glue. Heck even changing the layout of apps on my home screen alters what i engage with, and remember to do.

Now generating apps will be game changer. But i dont think it means apps go away. I think we just get better apps, and more open apis (for llms to use).

[stronglikedan]

> They know Spotify's algo can feed them novel music.

lol, hardly!

[yuppiepuppie]

> You don't need "apps", you need data feedable to LLM and a visualization toolkit for presenting results.

Unless Im missing something, how do you capture this data? It needs to be some sort of UI friendly method, as I dont see my mom chatting with an LLM all day about her dog's location (via GPS) or my dad counting calories through ChatGPT.

[rchaud]

It'll probably require access to your contacts, call history, messages, geolocation and browsing information at all times. Maybe even a Microsoft Recall-style data suck where it takes periodic screenshots of your activity and ingests it into its knowledgebase.

[prepend]

> On paper, this sounds amazing. Like "out of sci-fi books" amazing.

How does this sound amazing? What about the specs amaze you?

This seems like every laptop currently made with a browser open to chatgpt/claude/gemini.

How much does this cost? What’s the local compute?

Isn’t the best AI laptop just a $600 Neo? I’m not sure how this is superior and I just read a whole web page about it.

[jmye]

> The caveat, though? I very much doubt Google has the capacity to execute this properly.

I don't think it's a capacity issue, it's that their whole entire business is advertising and acquiring the user data to feed it. I don't see any reason they would want to execute on this in a way that is meaningfully beneficial for users.

I suppose, my own caveats for some of the responses suggesting people like being advertised to.

[nfw2]

It seems to me like most of the predictions that the interface of the future will entirely be agents on top of systems of record are coming from people who never considered hci as a serious field. You can't predict something that exists is going to go away (eg apps designed around to guide specific user journeys) without first understanding the reasons they exist in the first place

[vovavili]

I would be curious to see how OpenAI or Anthropic would implement this vision without the inertia of a ~200,000 person company and with all the enormous brain gain these two companies now have. A Linux-based, LLM-native phone/laptop hybrid is pretty close to being my personal dream consumer device.

[kspacewalk2]

Can I somehow strip out the LLM-native part and just have a polished Linux-based mobile device, without someone trying to control what I run on it? Because that would be my personal dream.

[vovavili]

I'd imagine that there would be little financial incentive to spend millions on UX/UI and polish for this kind of device without an LLM capacity.

[kspacewalk2]

I'm okay with capacity, and turning it on by default. I just want the ability to chuck it in the trash can if I want to, and use my device without unremovable apps and features.

[vovavili]

The Zed approach indeed is the only consumer-friendly way.

[rchaud]

Isn't that what the Framework laptop does?

[benoau]

There are rumours OpenAI is making a smartphone -

https://techcrunch.com/2026/04/27/openai-could-be-making-a-p...

[trizoza]

+1 on having little hope Google can pull this one up well. Recently my Android auto updated to use Gemini instead of their legacy voice command, which already sucked, and it got significantly worse, it's trying to play songs on Audible and Libro.fm at random, while I even have Spotify running. And I'm on Pixel 10 Pro! When asking to play a soundtrack of a movie, it starts to play random cover instead of the official thing, just no idea how it can be so bad.

[mrhottakes]

Maybe that's what you want, but why is your preference what "people" "need"?

[22ekeke]

Imagine writing with such confidence about a thing that is akin to google Glass - DOA.

[rtgfhyuj]

product dependent on LLM what could go wrong? a sucker born every minute