rbranson

tl;dr is that "ultra-processed foods" are engineered to be more palatable and that tends to result in more caloric density.

One of the only (the only?) commercial grade implementations was launched recently by us at PlanetScale:

https://planetscale.com/blog/announcing-planetscale-vectors-...

The list of circumstances for WAL-mode busy errors is in the doc linked by one of the posters above. It has nothing to do with transactions.

How does busy_timeout address a performance aspect of SQLite?

Yeah, this almost never happens in practice. It’s not even worth being concerned about.

WAL mode makes those redundant. Just use WAL mode.

This is a really long blog post to just say that you should turn on the WAL if you want concurrency out of SQLite. All the other stuff is superfluous.

Yeah calling the EC2 API is definitely more complex than leasing datacenter space, purchasing racks of hardware, deploying a fault tolerant and secure network, capturing and managing offsite backups, dealing with hardware component failures, etc.

Incorrect. See: https://www.levels.fyi/companies/facebook/salaries/software-...

That's how risk works. The FAANG employee friends have exactly a 0% chance at a 9-figure outcome. They'll easily be at top decile if they pull a nominal $10M+ post-tax in 20 years.

The fact that New York (a city of 8M people) has only received ~150% more funding than Mountain View (a city of 81K people) tells you all you need to know about the promise here.

The reason is that runtime design forms the constraints of the GC design. Java allows relocating objects, Go does not. Go allows interior pointers, Java does not. These decisions have deep impact on GC design. Something this performance critical needs to be tailored to the runtime.

While I can't vouch for Ben's time at Airplane, I worked adjacent to Ben at Segment and can say his reputation there was unimpeachable and his work was top notch, so it doesn't surprise me that he'd publish something so thoughtful and informative.

Many people's biggest issue with BSL is that there are dozens of versions of the "additional use grant" which each have bespoke language with very critical clauses, none of which have any case law behind them (correct me if I'm wrong).

Even though software may be licensed under "BSL" it isn't really a standard, even though proponents tend to use the term "BSL" as if it is the same as talking about "GPL" or "BSD". The MariaDB BSL used here is quite different than the HashiCorp BSL, for instance.

It's awfully similar to a lane of PCIe 4.0, if we're talking about instance storage. It reads like they've chosen to map each physical device to a single PCIe lane. Surely the AWS Nitro hardware platform has longer cycle times than PCIe. Note that once the instance type has multiple block devices exposed (i.e. im4gn.8xlarge or higher), striping across these will reach higher throughput (2 devices yields 4G/s, 4 yields 8G/s).

Which SF program are you speaking of that continues to provide direct cash payments?

EBS attaches a single block storage volume to a single host[1]. S3 Express is a service-based object store. Apples and oranges.

[1] Yes, I am aware of multi-attach but this introduces a scaling bottleneck and requires a fairly exotic setup.

It’s probably more accurate to say that what your brain is doing when writing code is more like emulating an LLM than an LLM is emulating your brain.

Converting human language to code is an ideal task for LLMs, which is part of why they are such an exciting development.

Given how these blog posts are typically written, it is very likely the 1.2 million QPS figure is an all-time peak, not anything like an average.

PON is what's allowed fiber deployments out for consumer affordable prices and continued scalability of endpoint bandwidth, so "stop-gap" seems a little bit of an extreme position. There are a lot of environments where just the space and power requirements of a 10G port for each customer make it a non-starter, let alone the cost.

You're always sharing bandwidth. It's not like you have committed capacity end-to-end to every site on the Internet.

There's a bias where most of the content is produced by the most hardcore group of people. This is, of course, no different with home automation. You can do Home Assistant without making it your hobby, it's just that most of the posts you see are from people who are way too deep into it.

Outside of a few household-name Champagne houses, most expensive wine is expensive because it is desirable and the vineyards used to produce it are tiny. Standard economics.

That is how prices work regardless of industry or product.

Fauna is really cool technology, but I'll note that it lacks interactive transactions as a core design constraint. That does enable a bunch of nice things. However, that means users lose one of the most useful parts of transactional SQL -- simple, safe read/modify/writes.

The only way is to ship all of the "modify" logic into the database via FQL. Given how many times I've seen code that does this unsafely with SQL, probably safe to say that if the bar is raised to "rewrite the logic in a foreign query language" developers will opt for thoughts and prayers instead.

If you're asking about secondary indexes, it was just seen as a "later" feature we'd implement as a follow-up. It was definitely asked for, just never prioritized before I moved off the project.

We replicated our MySQL database to a SQLite edge at Segment in ctlstore: https://github.com/segmentio/ctlstore

We considered tailing binlogs directly but there's so much cruft and complexity involved trying to translate between types and such at that end, once you even just get passed properly parsing the binlogs and maintaining the replication connection. Then you have to deal with schema management across both systems too. Similar sets of problems using PostgreSQL as a source of truth.

In the end we decided just to wrap the whole thing up and abstract away the schema with a common set of types and a limited set of read APIs. Biggest missing piece I regret not getting in was support for secondary indexes.

Peak small-window nits is not the proper Rtings metric for viewing in a bright room. It's Sustained 100%, which is only ~300 nits on the brightest, most expensive OLED panel on the market. A bright LCD panel puts out 600-800 nits here. This is the closest metric to something like watching sports in a well-lit room.

It is already broken down block-by-block. See https://broadbandmap.fcc.gov/home