EPTOn November 11 2026 05:33 Jankisa wrote:
I really wouldn't agree that AI has no moat, in fact I think that is what Antrophic is counting on, courting the high paying enterprise customers and making their services sticky, and it makes a lot of sense.
Yes, you can, at any point ask your AI instance to export everything it knows about you, your habits, the way you operate and all the knowledge it gathered over the last x months or years you have been using it, but the export is never going to be perfect, it's always going to skip a bunch of things simply because even the best frontier models can't really add EVERYTHING to their context and spit it out, intentionally or unintentionally, that makes the move quite tricky.
I did a migration for a small(ish) for USA standards company, they basically exfiltrated 50 of their 250 employees and spun them off into a separate company, I was consulting on their email migration, by far the most tricky and the task they were most nervous about was the auth migration so they can all keep their Claude instances intact, programmers like to "talk to the same guy" they got used to, and from my workflows, it makes absolute sense.
Not only that, the interface, way of doing remoting, agentic workflows, hell, even design and style that Claude can pick up can be very sticky and make companies and individuals very hard to part with, so, basically, that's your mote.
In general, I think it's going to be like "I'm a Mac, I'm a PC" moving forward, you'll have your walled gardens and people will continue integrating more and more with their AI's.
Google's going to get a lot of the Gen Z / Gen A , I've seen it out there, any question you ask them and anything you tell them and they don't know if they should believe you the phones are out and asking Gemini. It's also much, much better integrated with Android phones then anything else, so long term I think that's their walled garden, it knows an average user way better then any of the 2 startups.
Anthropic will get most of the Millennial / professional market and ChatGPT and Meta will split the GenX and Baby boomers, depending on which one got to them first and how plugged in into Facebook they are.
I really wouldn't agree that AI has no moat, in fact I think that is what Antrophic is counting on, courting the high paying enterprise customers and making their services sticky, and it makes a lot of sense.
Yes, you can, at any point ask your AI instance to export everything it knows about you, your habits, the way you operate and all the knowledge it gathered over the last x months or years you have been using it, but the export is never going to be perfect, it's always going to skip a bunch of things simply because even the best frontier models can't really add EVERYTHING to their context and spit it out, intentionally or unintentionally, that makes the move quite tricky.
I did a migration for a small(ish) for USA standards company, they basically exfiltrated 50 of their 250 employees and spun them off into a separate company, I was consulting on their email migration, by far the most tricky and the task they were most nervous about was the auth migration so they can all keep their Claude instances intact, programmers like to "talk to the same guy" they got used to, and from my workflows, it makes absolute sense.
Not only that, the interface, way of doing remoting, agentic workflows, hell, even design and style that Claude can pick up can be very sticky and make companies and individuals very hard to part with, so, basically, that's your mote.
In general, I think it's going to be like "I'm a Mac, I'm a PC" moving forward, you'll have your walled gardens and people will continue integrating more and more with their AI's.
Google's going to get a lot of the Gen Z / Gen A , I've seen it out there, any question you ask them and anything you tell them and they don't know if they should believe you the phones are out and asking Gemini. It's also much, much better integrated with Android phones then anything else, so long term I think that's their walled garden, it knows an average user way better then any of the 2 startups.
Anthropic will get most of the Millennial / professional market and ChatGPT and Meta will split the GenX and Baby boomers, depending on which one got to them first and how plugged in into Facebook they are.
You do know that everything the AI knows about you lives in text files (or in the cloud, more realistically some cache db like redis)? It can just export them. No need to load them in context and regurgitate. If you migrated a bunch of Devs and they didn't know where to find their project and user level .claude folders I also don't know what to tell you.
Anyway, comfort is a shitty moat. Pycharm doesn't have a moat because I prefer it to VSCode. When all the AIs half-ass their integration with Pycharm and it all works smoothly in VSCode, I switch, even though there are things I really really miss and wish I didn't have to use VSCode. And VSCode is the umpteenth IDE I use.
I wouldn't even call AWS a moat. Their ecosystem is great, and it is a considerable cost to pick up and switch to Google Cloud, Azure or some other solution. But it's self-contained: I can't think of a specific service in AWS that isn't available in other cloud solutions. They might work slightly differently and require some changes to work smoothly. If you pressed me, Zeppelin on Cassandra is probably the service that was hardest to get working in GCP. A reverse migration would probably require serious compromise for Bigquery solutions. But even so, not really a moat: if Amazon felt the real need to develop a Bigquery-esque solution, they would do so quickly. Unlike Mac vs PC or iPhone vs Android where not only do you have the ecosystem differences, you have different hardware, and app purchases that don't transfer between ecosystems even if the apps are multiplatform, making it a much harder process than just what OS you prefer. Social Media has a giant moat: the users themselves create the moat: switching apps cannot be a solo action.
