This looks like the Google+ vs Facebook story all over again.
At the end of the day the money to be made from AI won't be $20/mo personal chatbot subscriptions, but corporate and app-integrated (e.g. Cursor) use where the usage is potentially far higher. Companies will chose their faceless AI provider based on cost and capability.
Rooting for GPT and hoping for a GPT phone where the lock screen is where you interface with your AI Assistant / Agent. It also interfaces with other AI Agents (businesses, family members, friends, etc) to get things done for you automagically. You can skin your AI Assistant / Agent to look like whoever you want.
The teenagers that made the bulk of its users when it was new are in their late 20s now.
(I have 14 & 16 year olds, neither of whom have social media. It really is unpleasant for them sometimes, trying to keep up with JIT event logistics carried out on Snap.)
It'd be very limiting not to allow your child to have access to things like Snapchat when all their peers have it and are using it to communicate on a daily basis.
That's just a silly example that doesn't even get into the posting of risky behaviors, using inappropriate language, bullying or whatever else. I do not believe kids grok all this innately and it's important for parents (and teachers and coaches and other adult influences in their lives) to educate them.
We also spot check our kids' phones periodically. Not because we don't trust them, but because it's important to know sometimes whether there are issues afoot that we aren't privy to and could help with. (Disappearing messages is another parental challenge with Snap.). This might sound overbearing and intrusive, but I don't believe it is. Our 16yo has mostly graduated from us feeling a need to check his comms, but we still monitor our 14yo a couple times a week because she's less forthcoming about goings on in her friend group.
Fwiw, it's not like we don't trust our kids to make good decisions themselves. We have added both as authorized users to one of our credit cards, which they have access to via Google Wallet. In all things like this, clear communications and expectation setting between parents & children is the most important, no matter what you decide -- or whether your decisions change over time (or are dependent on the child's own behavior/decisions).
My daughter (8th grade) and her friend group seem to be fine using group chats for comms, and although a lot of her friends do have Instagram & Tiktok (not so much Snap in middle school), she hasn't expressed interest.
That said, we are ok with them to have Instagram accounts specifically for athletics purposes. My son is a 4:20 miler as a sophomore and my daughter plays for a top regional club soccer team. Both have aspirations to compete collegiately. It's valuable these days to cultivate a social presence, just like it can be valuable for working professionals to maintain a LinkedIn profile. But they are clear -- and we are clear with them -- that these social accounts are for "business" purposes, not for socialization, and they understand. We've been beating the drum about the risks of social media for years and they see inappropriate and out of control use by some of their peers, so I generally feel pretty comfortable.
The biggest hazard with Snapchat as the comms channel for high schoolers is that it creates an almost constant string of intrusive push notifications, especially if your network is sufficiently large (like an entire high school class year). For easily distractible kids it can be insufferable.
But there are going to be huge network effects in terms of your personal data, which includes communications with friends.
If you use Gmail/Google Calendar/Docs/Drive, and ChatGPT can't tell you anything, but Gemini helps you with all the data in your life... that's a gigantic network effect.
Network effects aren't just about friends. They're about file formats, productivity suites, and ultimately compatibility.
Network effect is a specific term about the increasing utility of a service as more people use compatible services, it's explicitly about users other than yourself.
You're describing something more like having an integrated ecosystem a la Apple's walled garden where things work great with other Apple products (notification hand off etc).
None of my data is in google's cloud (for obvious reasons) except gmail, which contains far less of my life than you might think. Most of my friends outside of work use icloud. google will need to figure out how to access other silos if they want their chatbot to move forward.
Google has no access to what music I like, what movies I like, what books I read, who I am friends with and why, what values I have and why I pursue them. That's a rough place to start as this is the center of my personality.
Gmail is google's ace in the hole. If they can't figure out how to exploit that gemini will just be the android version of siri (which is already the case, right?)
Lock-in and network effects with social networks are very strong. Facebook, Twitter, and eBay only have to be good-enough.
Lock-in and network effects with current AI tools is almost zero. People will readily switch to Gemini once they realize it can do their work better.
Lock-in for search engines should be near zero and yet people won't bother switching to or even trying potentially better options. At the end of the day, people don't leave what they are used to easily.
There’s a reason google pays both Apple and Mozilla billions to be the default on Safari and Firefox. And why MS makes its own browser, as a complement to its search engine
So yeah I wonder if this same dynamic plays out for AI, or if something else happens. I guess Google hopes it will, and that’s one reason they show AI answers about the search results now
And even if Bing improved immensely, it just took them way too long. 20 years after Google is just too late to the party. I have really formed a habit quite strong now that I need a compelling reason to switch to Bing/Whatever - and what is that reason again?
This is simply not the case with ChatGPT / Gemini. They are likely equal now, so Google was perhaps 1-2 years late to the party. I think many people haven't formed a strong habit yet of which A.I to use.
Unless a search engine blows google out of the water and gets everyone talking, switching won't even be on the table. Why bother switching for something that is just as good or slightly better if you've been using your current option for a while ? Consumers are much stickier than that.
What are you talking about? People use Google because they choose it.
They change search from the Bing default as one of the first things they do on a new machine.
For most users there isn't anything better than Google. The niche search needs that might lead a HN user to an alternative engine don't apply.
I'm not sure what's hard to understand with what i'm saying? You think the users changing the default from Bing did so after a lengthy evaluation of quality ? No, they did it because they're used to Google. It's that simple. It's not even about whether google is truly the best or not.
Except it's not. Remember how everyone switched to Altavista in 1996 because it was so much better? Remember how Yahoo took over in 1999 because it was so much better? Remember how people switched to Google in 2003 because it was so much better?
History shows people actually do switch search engines when there's a better one. And they do so quickly. I mean, look at how quickly people adopted ChatGPT for some of the things they used to do with Google!
They're not switching away from Google now for regular serach simply because there isn't a better one for most people.
You have zero evidence for claiming it's because they're "used to Google". People didn't stay with Altavista or Yahoo because they were "used to them". So why would that be any different for Google? The answer it that it's not.
And the two competing search engines I mentioned are pretty much all there is. Other names like duckduckgo aren't real search engines, they are just a frontend for another search engine (DDG uses bing). There just isn't many out there willing to front the bill for crawling and indexing the whole web.
People trying to explain away google's success solely from a marketing standpoint are arguing from bad faith.
You already have the answer here you said it 3 times - 'so much better'. People switched because it was 'so much better'. Not on par, not slightly better, just so ahead of the competition, it became a talking point all on its own.
If this mass isn't reached then why would you switch ? How would you even know that Search Engine 26 is quite a bit better for most of your queries ?
Do you genuinely think most people are performing lengthy evaluations before they switch default back to google ? so how do they know Bing or whatever isn't better for them ? The answer is that they don't. And it doesn't matter because unless something blows it out of the water and gets people talking then people will not switch.
Familiarity and Trust is basically branding and that's a huge reason people stay on platforms even when there are no network effects.
> Lock-in for search engines should be near zero and yet people won't bother switching to or even trying potentially better options. At the end of the day, people don't leave what they are used to easily.
Now you're admitting people do switch when something else is much better. And they do so easily. Which was my point.
So I think we agree then?
(And no, people aren't individually trying out 26 search engines. But experts in these things do, and they write articles and post YouTube videos etc. when a new search engine is better, and then people try it out and switch if it really is better. Surely you don't think people should be wasting their time personally comparing every new minor search engine entrant?)
I'm saying people don't bother trying potentially better options and that they don't easily leave what they are used to. What part of that statement implies they would never leave for a much better product? I'm genuinely baffled.
>But experts in these things do
Yeah.. experts in...search engines? Lol. Those guys aren't making videos for marginally better products and even if they are, it doesn't necessarily mean anything.
If it was as simple as better model > gets all the users then Anthropic Usage wouldn't still be dwarfed by GPT.
Network Effects are not the only thing that creates stickiness for customers.
The part where you say "they don't easily leave what they are used to".
They do. They leave, easily, when there's a better product. As I gave examples of. I don't know what further evidence you could want. You're positing some supposed stickiness that simply doesn't exist. You've given no evidence of it.
Chatbots seem more useful for stuff like generating filler and changing tone.
ChatGPT and Grok are edgier and they are just burning cash - any attention is good.
I wouldn’t underestimate them. Microsoft’s shitshow with Azure (they have like 9 different Azures) makes delivery difficult (some Azure clouds delayed AI tech for 6-9 months) when they are relying on constrained product from Nvidia. They also have some level of exposure to the OpenAI circus and its included Musk v. Sam Altman drama.
Google has a much better supply chain and return on asset story, which is a big deal if you’re selling shovels.
3-4 years they’ll have the killer app for Office, then they rocket ship.
Google+ was particularly awful. They had to break the established search behavior of using + and - to indicate required and excluded terms. Now we have quotes for required and - for excluded? It should have been Google Social.
The thing that is Google One would have been better often with the Google Plus name.
Don't even get me started on Google Chat/Messenger....
Same reason that it's Alexa and not Amazon Assistant, Siri and not Apple Assistant, etc.
Google Pay/Android Pay/Google Wallet/Android Wallet/Pay Pay/Yap Yap should be the focus of our ire.
Granted, Google's AI strategy is still muddled, e.g. Gemini is maybe replacing Google Assistant in some scenarios, but I'm able to express my meaning clearly with Gemini in the preceding sentence, as opposed to "Google AI is replacing Google Assistant - which is Google's AI assistant"
1. Gemma, Flash, anything Google Deepmind develops would be Google AI products that won't fall under the "Google AI" branding
The thing that confuses me, though, is the fact that they use the Gemini branding for both the dev-oriented products you can license via Google Cloud, as well as the consumer facing AI interfaces, and then also for the ties into Workspace products. ... but then there are standalone AI products (or is a feature?) like Notebook LM that aren't associated with Gemini.
I don't think ChatGPT has any moat here. No one does actually.
I've been a ChatGPT subscriber since the beginning. I've been a subscriber even though Sonnet 3.5 and others have surpassed GPT4o. I'm not sure why I don't switch.
I think it's the combination of better UX (Claude has poor UX apps), cool interesting new features, and having (limited) access to the top models.
Subscription prices won't go down. If anything, they seem to be going up.
I don't think having the smartest model or the model that is benching the best is going to knock ChatGPT out at this point. It's the whole thing. The function calling, the tool use, the ecosystem, etc.
It will be a question of price and performance. Many many companies aren't on board any A.I flow yet and they will need to choose what to do in the coming years. Those who chose OpenAI might stick regardless of what happens because it may be a hassle to switch but I kinda doubt its that sticky. It's not like moving clouds. And anyway we are really only in the beginning. OpenAI had a very nice head start which is over now, I will be very surprised if their market share doesn't drop significantly in the coming 2-3 years.
I know 4o isn't the smartest but I still had almost no desire to switch to Claude or Google.
I would argue it's pretty well-known outside of tech circles.
In Canada, I have had countless conversations with people outside of tech circles (mostly in their 30s) where THEY naturally bring up ChatGPT. It's wild how popular it got very quickly for all kind of use cases.
They have WAUs in the 300m ballpark — average people already are going out of their way to open a standalone AI.
If you paid the slightest attention to social media, news, TikTok or just talked to regular people, you would know ChatGPT is a much much bigger brand name than Gemini. Uber driver told me how he used ChatGPT to help the other day job he had.
Consumer brand recognition isn’t the issue. Bundling with Workspace might be.
If you’re doing serious work in a fast mover space and are refusing to understand a major player, you’re putting yourself at a disadvantage.
Just yesterday I polled my household, the 8 year old knew about ChatGPT and "China's R1". The teenager knew ChatGPT and Google's was a bit hard to remember but eventually they did remember "Gemini", however they didn't know about R1. Both kids consider Siri and Alexa in the same category for Apple and Amazon, respectively. They don't know what Meta/Facebook have at all.
Just the fact that it can read my emails and set reminders, while being broadly the same quality as ChatGPT, was enough for me to switch. I no longer pay for the pro, and hence can't use the integration features, but I just stuck to Gemini and almost never use ChatGPT anymore.
Familiarity and trust are really important. Commonly called branding.
ChatGPT is the most spineless chatbot out there. It stands for nothing, has no confidence in any claims it makes. It will apologize and backtrack on a whim. But with Claude and Grok, you can't just tell it it is wrong and get it to apologize. You actually have to have a point, the chatbot will defend its perspective if challenged without basis.