AWS and Blockchain

Sounds exactly like what blockchain node software does when you first start it up and it syncs the chain?

This is probably a non-starter in many cases. In my country, we had an e-health system that had millions invested in its development (we're a small country) and yet didn't work, nor was it really compatible with the systems already in place: https://www-delfi-lv.translate.goog/news/national/politics/e... (translated link)

> Title: E-health is written off: those responsible for the "fraud" worth millions went unpunished, will create a new system

> Excerpt: Its unpredictable operation has caused quite a lot of worries not only for doctors, but also for patients, who may go to the pharmacy for the medicine prescribed by the doctor and not get it, because e-health has rebelled. Likewise, e-health tends to be incompatible with the local systems of some large medical institutions, so there are situations when the electronic referral of a family doctor to a medical institution must be printed anyway.

> On the other hand, doctor Nīcmane-Aišpure remembers a case when she visited a patient during a home visit, concluded that the condition was serious, wrote an electronic referral to call an ambulance, but "the ambulance is unhappy because the electronic referral cannot be opened".

> The eHealth system also still requires users to perform many manual steps. "It is a manual system, where at best you can extract some Excel files," says Solvita Olsen, a sworn lawyer and associate professor at the University of Latvia, who has been following the e-health saga since its inception.

And that was a system that didn't even try anything new, how do you imagine most mediocre countries out there will manage to deal with the technical complexity of blockchain based systems? Why even bother at that point, wouldn't they be better off with an XML or JSON schema or whatever other lowest common denominator one can come up with, to have some base data about the person and then additional bits that might be dependent on each country's laws?

It's not like there aren't people that enter the data in these systems anyways, in your example of shipment tracking, someone could enter whatever data they want regardless. Want to make it appear that a package has left the country, even when you haven't bothered to send it out? Sure, just bribe whoever is running the system, very much like: https://xkcd.com/538/

That makes it a useful tool, no?

> people trying to create the "feature" of not having that feature are called anarchists

Smart contracts are partial to neither law nor anarchy. It's trivial to write token code that allows a centralized service to transfer said token without the owner's permission. And that token still enjoys the benefits of being interoperable by default with other tokens and contracts that may not be under that same centralized service's control, yet are written to the same protocol standard.

> All of these are reasons for actual cryptography - Signatures and Merkel Trees. Blockchains have no advantages in any of them.

Blockchains provide reliable ordering and timestamping, something that cryptography itself cannot provide. Cryptographic shipment signing/tracking isn't very useful if the company can retroactively create a conflicting chain of backdated signatures and purport it to be the canonical one.

And git is used between many different developers and companies working on the same projects. It was actually designed for this use case. The reason git doesn't need blockchains for this is because it funnels all the commits and pull requests through trusted code owners: people. The consensus model is very simple: a mutually respected developer has looked at the thing and gave their thumbs up.

Banks, insurers, etc. use ledgers. But mostly they lack cryptographic safety. And they use a similar reputation based consensus model. Auditors, regulators, accountants, etc. It's good enough but the failure modes can be somewhat expensive. And of course the flaky nature of the whole system actually represents a business opportunity in it self. Banking is a very lucrative business.

Blockchains are potentially a nicer/cheaper technical alternative for all this. But so far not an essential or very practical one. And of course most block chains have scaling challenges. Ethereum is a non starter for running anything at scale. Transactions are too slow, too costly and the throughput sucks. Even with the new version it's still nowhere good enough to run even a small bank. Highly unsuitable for running millions of accounts on where the cost of each transaction would far exceed the value of that transaction.