No, the author is not confusing them. It’s everyone else that has been confused insofar as they use “REST” to describe APIs that are not actually REST. True REST is a way of architecting networked applications according to a set of constraints, as defined by Roy Fielding in his original dissertation. Here’s an excerpt that sums it up:

In order to obtain a uniform interface, multiple architectural constraints are needed to guide the behavior of components. REST is defined by four interface constraints: identification of resources; manipulation of resources through representations; self-descriptive messages; and, hypermedia as the engine of application state.

Many “REST” APIs do not adhere to these constraints, especially self-descriptive messages and hypermedia as the engine of application state (also known as HATEOAS). Note that this says nothing about HTML: HTML is but one representation of hypermedia, but far from the only one. Hypermedia is something that allows discovery of additional, related resources by way of links between resources. There are a number of ways of doing this, such as RFC 8288 Link Headers or embedding links within one of a resource’s representations (such as a JSON representation).

It’s a much larger topic than I can fully explain here, but I highly recommend reading Roy Fielding’s dissertation, especially Chapter 5 which talks about REST and what it is.

I also highly recommend this blog post from Google Cloud’s engineering team, which talks about why you should use links instead of keys to represent relationships in API design: https://cloud.google.com/blog/products/application-development/api-design-why-you-should-use-links-not-keys-to-represent-relationships-in-apis.

From what I remember, Vscode has quite a fair bit of native code in various areas to make it perform better. So it may not be the representative Electron app for comparison.

That’s what Open Telemetry is for, honestly, specifically tracing. I’m in the process of adding tracing for all of our web handlers, http client calls, database transactions/individual queries, and asynchronous jobs. Should give us a pretty good view into the performance of the system, and allow us to monitor for things like performance regressions without any dedicated profiling needed (though we certainly could add additional profiling information through Otel down the road). Since I know you’re a Haskeller (recognize your handle), the hs-opentelemetryecosystem is quite solid and has integrations with a lot of libraries in the wider ecosystem, like wai, persistent, postgresql-simple, conduit, and a bunch more. I haven’t needed to write too much of our own instrumentation, aside from our job system since that’s a bespoke system.

We are fully capable of thinking PHP, NPM, and PyPi are trash simultaneously. Though it should also be said that only PHP has had decades of security flaws in the language itself, rather than a package manager.

It’s just a very poorly designed language with countless footguns and fucking idiotic standard library functions, on top of being a running joke, security-wise.

Or, you know, your middle schooler can just draw silly things like we all did growing up.

It’s not, they are correct. There are so many telltale signs of AI code it’s not even funny. The most obvious being the absolutely garbage comments that people invariably leave in. And the god awful naming of things.

But it’s more than that. If you’re in this industry long enough, you get pretty good at reconstructing a real author’s thought process when they wrote the code. Part of understanding code is understanding why it was written the way that it was, which means understanding how the author thought. Sure, there was plenty of garbage code and stupid thinking, but it was still thought and usually had its own kind of internal consistency. AI code has no thought process, no reason. It is pure mimicry, an amalgamation of millions of authors all at once. It has no internal consistency. It produces code no human ever would, because it doesn’t make any fucking sense. It’s like reading an alien language, where the constructs it produces are so aberrant to how any human would ever think that it could be nothing but an LLM.

I definitely can’t tell you everything about a piece of code by reading it, but I sure as hell can smell an LLM in it, because it fucking stinks.

I work remote, so no.

Hmm, every company monitors their web traffic. That’s standard operating procedure.

Lisps are hardly my favorite languages, but the whole point of s-expressions is that it is a uniform syntax, which makes metaprogramming much, much easier. Metaprogramming is kind of the whole thing of lisps.

Ah yes, the “gotcha” everyone loves to trot out. Except Lemmy is quite obviously not what anyone is talking about when they talk about the issues of social media.

Social media is definitely harming minors, just like it is harming everyone else.

I just get nonstop recruiter spam from Anduril to make murder robots, even though I have repeatedly told them to fuck off and take me off their lists.

Deadlocks aren’t generally too much of an issue with an MVCC DB like Postgres (though it certainly can happen if you use explicit locks and such), but yeah ultimately it’s all about the kind of workload we’re talking about. I honestly think of it much less like a database and more like a high-performance synchronization mechanism that can cross process boundaries. There’s many times where you want to coordinate the behavior of many independent processes (like horizontally-scaled services), such as rate-limiting your requests to an external service. Implementing a distributed rate-limiting algorithm with a RDBMS is a pretty poor choice, because the data used for any synchronization is ephemeral, unimportant, and short-lived (and this, ACID guarantees are pretty irrelevant), and you would incur a lot of contention overhead for shared resources. You might even get bottlenecks that prevent you from achieving the rate you’re trying to limit yourself to. Something like Valkey gives you simple, fast atomic writes with built-in TTL mechanisms, which is perfect for a rate-limiting system. There are many other examples like this where it makes a lot of sense.

Or consider even something incredibly simple like, I dunno, a realtime user counter. Valkey has an atomic primitive operation for this built-in (INCR). With an RDBMS, this is not trivial as you scale up the number of clients, because either you have to deal with a ton of row-level contention (if the counter is stored in a single row that gets updated), or you end up hammering your disk with increasingly large aggregate operations (if there’s a row-per-user and the total is computed across them, somehow).

As I said, it’s really just very different workloads where they are valuable.

I’m not sure what you mean by “standard DB queue” (perhaps you mean a message queue, I guess?), but yeah Redis (or really, Valkey, now, since Redis tried to change license on everyone) is a very important tool in the toolbox for system design. In-memory databases trade persistence guarantees for raw throughput, and are very well-suited for different workloads than a traditional database is. It’s great for caching, rate-limiting, realtime analytics, leaderboards… Basically anything with ephemeral data or data that can easily be reconstructed from some other source of truth. It’s an optimization tool that can significantly reduce the load on your primary application database for certain workloads (and achieve higher throughput).

Slop commenting. The litany of AI slop websites that have popped up, polluting search results. The absolute flood of AI slop pull requests being submitted to open source projects, to the point at which many open source maintainers are buckling under the weight of garbage being thrown their way. Bots everywhere. The list goes on and on and on.

The internet is already a dessicated husk of what it once was.

Like when my toddler says he hates a food that he has never tried.

I’m sorry to hear about your condition.

I’m saying that AI companies have repeatedly shown a flagrant disregard for licenses, contracts, and the law, and have engaged in massive theft. Why should we expect that any contract would actually be honored? The whole thing is a huge fucking grift.

Our interview process has the candidate implement a couple (extremely simplistic) API routes, with the specific approach left up to the candidate. It’s more about having something they can talk through and evolve than it is about us looking for any specific approach.

Basically it exists as a way for us to have a starting point for talking about scalability topics, asking the candidate how they would evolve their implementation as we scale up the system. Does a pretty good job of giving us an idea of how they think through system design problems.

I was not talking about your specific lawsuit, or TLS.

Companies trust other companies all the time, and it’s foundational to most all SLA/SLOs. Any time I’ve voiced concerns around how AI companies are using the data we are giving them (like giving them access to our codebase), it’s brushed off as “we have an agreement with them”. It’s just a load of hogwash. They can and will abuse all data they have access to, just as they have done thus far.

In this particular case, we are talking about a data center, and it is not at all reasonable to assume that the data that flows to said data center is in any way protected, especially one run by Musk.