4 ms·

The times for "resume driven development" are over, yes. And I'd argue that's a good thing - there just isn't any money left to develop "the next hot database"

The times for "resume driven development" are over, yes. And I'd argue that's a good thing - there just isn't any money left to develop "the next hot database" or whatnot, with hype chasers forcing developers to reinvent the wheel and make all sorts of basic errors that other mature tech has long since passed again. Maybe the industry can now focus on making the hype tech of the last 20 years finally somewhat stable and secure again before the next boom-bust iteration hits...

DeathArrow 2y ago
>there just isn't any money left to develop "the next hot database" Are or problems with distributed data and databases solved? You can make money by developing a database that solves some problems, saving time and money for lots of people and companies. "Hot" doesn't make money, otherwise people posting here daily "I rewrote X in Rust" would be rich.
- mschuster91 2y ago
  > Are or problems with distributed data and databases solved? Unless you're building the next Facebook or other billion users scaled service, mysql or postgresql will be more than enough for everything, at least until you hit 100k-1M users. And by that point you have enough real-world experience to accurately judge your needs and if you need to roll your own stuff. As for distributed data - RDS and a multi region database has that problem solved. Unless you're a real time bidding ad broker, your app can get away with a few milliseconds of replication latency.
- lelanthran 2y ago
  > Are or problems with distributed data and databases solved? What do you need it for? I'm not being facetious, I'd really rather like to know. What I see is that this sort of thing is only needed when you are planning to have the next facebook/reddit/etc type of business. I write line of business software for different SMEs, and I have not yet found a need for anything more than a beefy server/VPS. Even companies with 100k employees each using the same piece of software does not need anything "distributed", if you're even a little careful with read-only replicas, some caching of state, etc. You want to geolocate nearer to some users? For the reddit/facebook type of business, sure, good idea. For the majority of software, why would you? That webapp functions only slightly degraded when used from the other side of the world.
Apocryphon 2y ago
None of that is happening. The hype is now around generative AI, a prototype tech whose main benefits to the end user are still largely speculative. (At least with the gig economy, consumers can get cheap rides and food deliveries. At least with social media, they could share cat videos.) I am dubious that the industry is any more interested in making hype AI any more stable or secure than they did prior waves of hype tech.

DeathArrow 2y ago

>there just isn't any money left to develop "the next hot database" Are or problems with distributed data and databases solved? You can make money by developing a database that solves some problems, saving time and money for lots of people and companies. "Hot" doesn't make money, otherwise people posting here daily "I rewrote X in Rust" would be rich.

mschuster91 2y ago
> Are or problems with distributed data and databases solved? Unless you're building the next Facebook or other billion users scaled service, mysql or postgresql will be more than enough for everything, at least until you hit 100k-1M users. And by that point you have enough real-world experience to accurately judge your needs and if you need to roll your own stuff. As for distributed data - RDS and a multi region database has that problem solved. Unless you're a real time bidding ad broker, your app can get away with a few milliseconds of replication latency.
lelanthran 2y ago
> Are or problems with distributed data and databases solved? What do you need it for? I'm not being facetious, I'd really rather like to know. What I see is that this sort of thing is only needed when you are planning to have the next facebook/reddit/etc type of business. I write line of business software for different SMEs, and I have not yet found a need for anything more than a beefy server/VPS. Even companies with 100k employees each using the same piece of software does not need anything "distributed", if you're even a little careful with read-only replicas, some caching of state, etc. You want to geolocate nearer to some users? For the reddit/facebook type of business, sure, good idea. For the majority of software, why would you? That webapp functions only slightly degraded when used from the other side of the world.

Apocryphon 2y ago

None of that is happening. The hype is now around generative AI, a prototype tech whose main benefits to the end user are still largely speculative. (At least with the gig economy, consumers can get cheap rides and food deliveries. At least with social media, they could share cat videos.) I am dubious that the industry is any more interested in making hype AI any more stable or secure than they did prior waves of hype tech.