mborch

Deployments can be on-prem or cloud-based depending on needs and constraints.

Would be nice to see the actual wording in the cable, but I suppose Reuters are not allowed to publish that; we get a cable paraphrasing a cable.

If batch sizes are sufficiently large, for example by staging updates, is this really necessary to achieve good insert performance?

Google claims "On average, EV batteries weigh around 454 kg" ... pretty heavy stuff to carry around.

Yes? https://kubernetes.io/docs/concepts/workloads/controllers/cr....

This 2011 addition to the XZ Utils Wikipedia page is interesting because a) why is this relevant, b) who is Mike Kezner since he's not mentioned on the Tukaani project page (https://tukaani.org/about.html) under "Historical acknowledgments".

https://en.wikipedia.org/w/index.php?title=XZ_Utils&diff=pre...

Arch Linux played an important role in making this compression software trusted and depended upon. Perhaps not a coincidence, but at the very least, such a big project should more carefully consider the software they distribute and rely on, whether it's worth the risk.

Here's another reference on this: https://www.yoursoundmatters.com/vinyl-vs-cd-in-the-loudness....

> As explained earlier, due to the physical limitations of vinyl, there are limits as to how loud you can press a record, and because vinyl is “for audiophiles” – there is less incentive for record companies to compromise the quality of vinyl releases. As a result, many vinyl records are mastered differently to the CD release with more dynamic range and at lower volumes.

But I read some place that Radiohead themselves preferred the more compressed sound, perhaps owing to listening in a car rather than a high fidelity setup.

Ah of course, that makes sense, thanks. In their comparison, perhaps they should still mention that just to clear up any confusion.

They mention https://pola.rs/ and highlight how it has the exact same properties (out of core etc) but don't include it in benchmarks and no mention on their detailed comparison page.

The filmed entertainment industry is pretty massive and AI is going to make a big impact there (for better or worse), that's probably something most folks can understand. CGI is already used everywhere and generative AI takes this trend much further.

ChatGPT knows how to answer this question, but how I don't know. Perhaps it's programmed to answer that question?

On MacOS, you have built-in sandboxing via "sandbox-exec" which shouldn't incur any noticeable overhead. It's used by Chrome, Bazel, etc.

Not sure what's available on Windows.

> A program run with Deno has no file, network, or environment access unless explicitly enabled.

You can do this using containerization technology, no need to invent this per language runtime.

They've "invested 5 million engineering hours" which is impressive, but are the design changes impossible for Apache Kafka to adopt?

Since any argument after one or more "**" must be a keyword-argument, perhaps f(pos_a, pos_b, **kwargs, kw_a, kw_b, ...) or even f(pos_a, pos_b, **, kw_a, kw_b, ...).

This would make the proposal's first example read:

    my_function(**, my_first_variable, my_second_variable, my_third_variable)

More readable IMHO. They do mention f(a, b, *, x) in the proposal and present some objections, but arguably ** makes it easy to understand that these are keyword arguments.

Seems like the two main difficulties in moving forward on this proposal are the name itself and agreeing on an API that's idiomatic to JavaScript, basically finding the right balance between simplicity and performance.

Perhaps a case of perfect being the enemy of good here.

Perhaps this should be an incentive for library authors to finalize the 1.0 API and accept that there's eventually going to be a 2.x down the road, a "less is more" approach.

Perhaps their team could have found something useful in KalDB from Slack which is based on Lucene:

https://github.com/slackhq/kaldb

Looks like you're asking for syntax sugaring for T::new wrapping, e.g. `"bar" |> Something |> Arc |> Mutex`.

Active support ended a year ago (12 Sep 2022), but somehow it still saw a security release on 24 Jul 2023. Perhaps we will see one more?

I have never met a productive software developer that didn't fundamentally enjoy his or her work ("happiness"). What do you do with an unhappy person? At some point, if everyone else is happy, it's just time for them to move on I think, and you probably have to "help them out".

You don't need to measure software developer productivity. The article (or ad) admits as much:

> Further, attracting and retaining top software development talent depends in large part on providing a workplace and tools that allow engineers to do their best work and encourages their creativity.

So do that!

Happy people are productive, and talented, creative people move your business forward.

I'll add that at the end of the day, a maintainer of open source needs to tend to its garden in order to receive contributions. This is a case of HashiCorp failing to do that on multiple levels (devoting too few resources, changing licenses, etc). So people leave to form new communities and that's only natural.

There's an article in American Journal of Men's Health:

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9134445/

> In summary, semen quality in men has declined significantly over the last 80 years, with sperm concentration dropping to approximately one seventh of its original value. This timing coincides with the development of plastics.

Not a bad hypothesis.

Don't they have some lock-in with "saved cards" – ?

JSON has pretty limited choice of data types though. Want better precision for numbers, use a string. Want dates? Use a string. Etc.

The meaning is pretty clear: you don't manage compute, it scales up elastically based on demand, even all the way to zero. Ideally, it reacts quickly enough to changes in demand that you don't need to worry about it. Serverless is basically the original promise of the cloud.

The race between Delta Lake and Apache Iceberg is a throwback to VHS/Betamax days.

Microsoft is betting on the former while for example Snowflake is now allowing you to back tables with the latter, making it possible to integrate your data directly with other systems.

I talked to a Microsoft rep who said that they would support Iceberg if there was a Rust-based client for it.

Interesting to see how Databricks reacts to this given that they have their own Project Lightspeed (replacing the Spark execution engine).

The challenge for Databricks and Snowflake is that big cloud vendors are catching up with their own serverless products such as Redshift Serverless [0].

And they now provide Spark engines as well with associated notebooks. What's the unique offering in Databricks seen in this light?

[0] https://aws.amazon.com/redshift/redshift-serverless/