> I am still unable to alleviate my scepticism that this entire space is vapourware, there is nothing here more than you can get with traditional database technology?
Technically, there could be some value in an append-only datastore that can be distributed - such as verifying where some product has come from throughout its supply chain and tracking evolution of various data series over time. Of course, there's nothing to prevent someone from entering malicious data into the chain anyways (e.g. if people did that with COVID vaccinations and recorded someone as vaccinated after taking a bribe, who's to say that someone couldn't do that with "ethically" sourced coffee, for example).
> At the same time, the amount of coverage and attention it is getting and for a sustained period makes me believe I am missing something.
We largely live in a post-factual world, so as soon as there is some supposedly advanced technology (the blockchain, machine learning, AI, serverless, Kubernetes, a while ago cloud computing also fit the bill) that nobody understands all that well, people tend to conflate speculation with fact and claim that every potential problem could be solved with this technology, oftentimes to extract resources from potential investors or other people with decision making powers (e.g. software engineers striving for resume driven development).
To those people, it doesn't actually matter whether the technology will solve the problems or actually even work, even the people that believe that these things will work oftentimes only do so due to deluding themselves, because it's not like you can easily test those claims.
In my eyes:
- the blockchain (in any its form, excluding event-sourcing which is an architectural pattern and basically a generalization of the approach) isn't preferable to a MariaDB instance most of the time.
- machine learning is good for a few problems but in many cases linear regression and simple algorithms are enough
- we are decades if not centuries away from "true" generalized AI, most of the current progress is simply research, not products that solve complex problems, use cases will be limited for now
- serverless is still someone else's machine, there is merit to the approach, but it's not suited for all systems
- Kubernetes is cool when you have the resources to support it, which is more than many companies out there are willing to spend (many don't even need it)
- cloud computing is cool for certain systems but in the end just boils down to the CapEx/OpEx question, it's just a different set of tradeoffs
- you can apply the same reasoning to other advanced yet hard to "test" technologies, like full self driving cars and the "hyperloop" and other concepts like that
I'm reminded of the Gartner hype cycle:
https://en.wikipedia.org/wiki/Gartner_hype_cycleIn our current world, it's really easy to end up in the "Peak of Inflated Expectations" which will also make plenty of millionaires out there, which also should be recognized - so at worst, it doesn't hurt to explore even the improbable technologies and maybe work for a well funded startup every now and then, unless you're not about that life. You might not always solve all of the problems that you chase after, but the money in your bank account will still be very real.
At the same time, it kind of hurts to see speculants get rich off of projects like "Earth 2" and empty promises.