Links

Century-Scale Storage

What would you use to keep (digital) data safe for at least a hundred years? Maxwell Neely-Cohen looks at all the factors, possible technologies, social and economic challenges that you have to contend with if you intentionally want to store data for a century. He explicitly chose that time scale, because it is at the edge of what a human can experience, but it is outside of a single human’s work life as well as beyond the lifetime of most companies or institutions. So the premise sets you up for a host of problems to be solved. He also analyses strategies for recording and keeping data past and present and evaluates their potential for keeping data safe at century-scale.
It’s long, but worth it.

We Don’t Want “Privacy”-“Enhancing” Technologies in Our Browsers

The current trend for privacy-enhancing technologies for surveillance in web browsers are going to be remembered as a technical dead end, an artifact of an unsustainable advertising oligopoly.

Don Martin has 10 succinct points on why users (aka we) don’t actually want so-called Privacy Enhancing Technologies (PET) … some technical, some social, some economic.

Best “AI”-Rant

Most organizations cannot ship the most basic applications imaginable with any consistency, and you’re out here saying that the best way to remain competitive is to roll out experimental technology that is an order of magnitude more sophisticated than anything else your I.T department runs, which you have no experience hiring for, when the organization has never used a GPU for anything other than junior engineers playing video games with their camera off during standup, and even if you do that all right there is a chance that the problem is simply unsolvable due to the characteristics of your data and business? This isn’t a recipe for disaster, it’s a cookbook for someone looking to prepare a twelve course fucking catastrophe.

How about you remain competitive by fixing your shit? I’ve met a lead data scientist with access to hundreds of thousands of sensitive customer records who is allowed to keep their password in a text file on their desktop, and you’re worried that customers are best served by using AI to improve security through some mechanism that you haven’t even come up with yet? You sound like an asshole and I’m going to kick you in the jaw until, to the relief of everyone, a doctor will have to wire it shut, giving us ten seconds of blessed silence where we can solve actual problems.

After some general ranting the author answers several common “reasons” why a company might want to use LLMs/AI tools.

Usefulness of Swap Explained

Chris Down explains how swap’s main role is being the missing backing store for anonymous (i.e. allocated by malloc) pages. While all other kinds of data (e.g. paged-in files) can be reclaimed easily and later reloaded, because their “source of truth” is elsewhere. There’s no such source for anonymous pages hence these pages can “never” be reclaimed unless there’s swap space available (even if those pages aren’t “hot”).

Linux has historically had poor swap (and by extension OOM) handling with few and imprecise means for configuration. Chris describes the behavior of a machine with and without swap in different scenarios of memory contention. He thinks that poor swap performance is caused by having a poor measure of “memory pressure.” He explains how work on cgroups v2 might give the kernel (and thus admins) better measures for memory pressure and knobs for dealing with it.