... while at the other end, howeer, most of the AI stuff still is owned by huge corporations that are able to afford the hardware and operation skills to handle AI. In my environment, this has even more pushed moving to the cloud, most specifically Azure as they also host GPT and similar models in European datacenters which eases pain of those who always wanted to take that step but feld hindered so far by GDPR obligations.🙈
Think about it this way:
There used to be one or two ovens per village, either at the bakers or in the town square. It took hours to heat up, by packing it full of precious wood, so everyone would get their bread from there, as building an oven and burning all that wood, just for one loaf of bread, would have been crazy. There were usually only a couple of kinds of bread being baked, as the work around the oven was so laborous.
Then people eventually got wood-burning kitchen ovens, that doubled as a heater and a stove. Then they baked their own bread, since the time and cost became negligable. But Most people Just baked the same bread over and over. At the most, they'd break it up into dinner rolls, or something.
And then everyone got electrical ovens and it got even easier and cheaper... and they all stopped baking. Now, almost everyone goes to the grocery store or the master bakery, where there is a wide, curated selection of bread, and choose from that selection.
In the same way, AI app generation won't lead to everyone creating their own apps. It'll lead to people becoming app-generation specialists and they'll offer an array of apps to choose from and further customize, hosting options, services, etc. Or they will concentrate on the data and storage layer, where the value is.
The dev talent will have to go one level deeper, in multliple directions, same way that baking had to become a factory good or an art form.
That is the effect of commoditization: go upscale or downscale.
I agree. I wanted to add that, in the future, most people will be willing to pay directly to client entrepreneurs and other service providers because doing so will overall improve their opportunity costs
Exodia38
Ok, now I understand, thanks. My opinion about this topic is that in the future everyone will be able to make clients with pretty little investment (almost zero). If clients become increasingly interoperable over time and everyone can create them, this means everyone has individually the benefits of interoperability BUT that doesn't mean those clients won't come without tradeoffs (see, I'm one of those that think that everything comes with tradeoffs. Also, the more useful or convenient, the more tradeoffs it has)
If everyone has to deal individually (or even in a Dunbar number fashion) with those tradeoffs, that means everyone will have to soften them paying MONEY (summing it up: the fact that everyone can make highly interoperable clients almost for free means they will have a clear DEMAND for the things that make clients smoother). Another thing is that individuals won't be really able to deflect the cost of making clients smooth by passing that responsibility to big corporations because that will be economically inviable in a free market
Everything I said makes me conclude that there will never be free lunch again because clients will be an eternal source of demand for everything Nostr related (Relays, dev work, etc) and the starting point for a sustained division of labor.
So, saying that relays are cost centers is highly gauche
View quoted note →
