[basch]

The thing still giving me pause is a lack of "bring your own model connection" between saas uh services.

If I already pay for Gemini Advanced (or OpenAI/ChagtGPT Pro), I then have to pay for it again at every service that offers a Pro 2.5 (or 4.1/4o) tier. I should be able to connect my Gemini Advanced access to any service that offers Flash and be able to upgrade. Signing up for a bunch of services is starting to feel like being triple, quadruple or more dipped. Similar to how I am annoyed seeing media content cross licensed to three streaming services and not getting a bill reduction when subscribed to multiple services with the same content.

[eightysixfour]

I really feel like the software model is going to start to break in a couple of ways in the near future.

Instead of vertical software slices, many people are going to want their single “horizontal” agent that they pay for (e.g Gemini Advanced, Claude) and connectivity to all of their other services.

MCP (which I personally think is a mediocre protocol, so it will probably win) as the glue for a bunch of services we OAUTH against and choose when our agents do/do not have access to certain tools.

The idea of a “GPT” App Store from OpenAI was sort of right, but just wrong enough. We are going to have an App Store inside of our preferred AI platform and subscribe/connect our other services from there.

[basch]

I prefer identity as the top layer. Give me an identity portal that acts as an account manager, access to api layer, let me perform password resets, and let me avoid 2fa if I’m logged into my identity manager account. It should allow bidirectional control: both what services have access to my identity, and what services anything in my identity are granted access to access. Things like LLM, cloud storage etc should be obscured away from services so they can’t tell my storage providers from one another. All access between services should cascade down from my identity portal.

[orand]

Who is going to pay for identity as a product and not just expect it as a free feature of another product suite?

[basch]

Why not both? Use Apple or Microsoft or a third party identity provider.

[jcgl]

I mean, that sounds basically like Oauth provided by the LLM provider, right? The provider just needs to build out the features.

[siva7]

No question, the openai gpt marketplace/store was a birth failure and doesn't make much sense in hindsight anymore. I just don't understand why they don't pull the plug and admit it's a failure.

[heywoods]

Your intuition regarding the shift from vertical to horizontal integration is spot on!

Sam Altman, in a recent Stratechery interview, detailed parts of OpenAI's future strategy that align with your prediction — a persistent, personalized AI. He envisions users interacting with OpenAI not just through core products but also across other applications.

Altman described a key part of the strategy: "...we have this idea that you sign in with your OpenAI account to anybody else that wants to integrate the API, and you can take your bundle of credits and your customized model and everything else anywhere you want to go".

This system aims to create a portable AI experience and by virtue, would usurp the vertical software business model that has historically dominated the software economy. A horizontal play, that sits in the middle collecting their tidy sum of the pot will require a very compelling argument. That would require a low barrier to integrate for developers coupled with a value-add proposition that is meaningful and not possible for anyone other than the largest technology companies.

As you know, it’s sort of the Wild West of tech right now. OpenAI is looking to find a territory in the AI landscape and make their stake now, and I think this is the correct strategy. We have seen what being the first to market with a great product can do for the longevity and growth of tech companies - especially the consumer markets. They have the name recognition, forever embedded in the lexicon of the internet, and a great product vision that will lead to critical mass adoption that and what awaits them is the coveted moat, at least in the consumer market, that AI companies have been struggling to find out in the Wild West of AI.

Altman mentioned wanting users to "be able to sign in with your personal AI that's gotten to know you over your life". This sign-in would ideally carry "your memory and who you are and your preferences and all that sort of thing" across different integrated services.

The OpenAI SSO login will be the Trojan horse and later on the app developers will either be incentivized by OpenAI or compelled to integrate their products because of the compelling value proposition it would bring to bare with an integrated personalized AI assistant, complete with its memory and preferences.

Lastly, I suspect this is one of the driving motivations to become a consumer hardware company as there is little to no chance that current players (Apple, Google, Meta) would allow the same 1st party access to their internal API’s would be a requirement for what Altman has laid out for OpenAI moving forward.

[freehorse]

I understand we some may want that, but why should everybody else want that? I operate in several contexts. I want some kind of models in one context and one in another. I want the AI to be contextualised within each use, not “personalised” on me. Yeah sure it would be great if it had memory within some project, but I would rather not have memory between projects or use cases that are irrelevant to each other.

[esafak]

The industry solution to this is to create an open standard for AI memory.

[everforward]

I doubt that fits in the usage model for those pricing plans. They're priced assuming it's inconvenient and less used because of it. I keep hearing they're losing money anyways, so despite being triple or quadruple dipped you're still paying less than it costs to run the models.

[basch]

How would it cost Kagi anything, if I bring my own keys and all the costs are offloaded to my personal Google/ChatGPT account? I pay them the $10/month and get the Unlimited Pro access (but only to models I subsidize the cost of.) If anything it would save them money to offload my usage to my account vs using up some of their Flash tier quota.

It would be ideal for them for LLM access portability to let them offer higher end models and have the end consumer pay directly for the usage.

Id much rather pay for one AI License, and a small fee to each service I use, rather than paying a high tier AI Bonus price at every single service.

[macrocosmos]

I think they might be talking about the cost to the model providers.

[basch]

Then Google should offer Advanced Portable for $30 that includes being able to plug it into third party services. With some kind of policing to prevent it from being used as the entire backend to a service (for all users) itself.

[everforward]

Third party usage is too disparate for a single flat rate to make sense. Perplexity probably uses orders of magnitude more tokens than Kagi does (the former uses tokens on every search, the latter only does if you manually invoke the Assistant).

Then there's the disparity in how many third party services each user uses. Some will use one, some will use 50. Google could charge per integration, but that's basically the status quo.

[basch]

Why would the number of integrations be the price point and not token tiers? I might want to connect my account to 100 services to test them, but only perform a couple uses each before forgetting about it.

[yellowapple]

> I keep hearing they're losing money anyways

Hearing from whom?