arionmiles

When I first started reading seminal works in Distributed Computing, I picked up the original (The Part-time Parliament) and it started off pretty well with the made up story of finding the algorithm as a manuscript from an archeological dig.

The original paper then quickly becomes super convoluted by continuing to explain the algorithm by overextending that allegory. By the mid point of it I felt pretty exhausted.

It was then I learnt that the original paper was lying in limbo for nearly a decade until it was finally published.

Then at a conference, Lamport got tired of people telling him his original was difficult to grasp and so came this simplified explanation.

I still have my notes on the paper somewhere in my Obsidian. I should publish those.

AI slop comment

I for one, am absolutely fascinated with Tron Legacy. It was the first Tron movie I saw as a kid in middle school. In some ways, it's responsible for the trajectory of my career.

Apart from the obvious reasons about the DP soundtrack and the visuals, I love the theme of chasing perfection and the way it backfires.

Kevin Flynn says to CLU in the end "The thing about perfection, is that it's unknowable. You don't know that because I didn't know it when I created you" and I love the fact that it says how we can put our best and our worst into what we create. That we're not just responsible for lifeless machines, that it's more than that. And it's a hauntingly beautiful thought.

Dwarves of Moria get swole

sure.. WWDC also has "Developers Conference" in its name.

Watching these bland presentations with choreographed delivery and reading off a prompt off-screen (I'm not completely sure they're doing this, but it looks like it) makes me appreciate Steve Jobs presentations from the past so much more.

Steve really had product presentations down. I wish people at least tried to copy him.

He just... highlighted Avatar. He clicked the dropdown menu, and then he randomly selected Papyrus. Like a...Like a thoughtless child just wandering by a garden, just yanking leaves along the way.

I was going for a "they performed surgery on a grape" thing but it seems I missed the mark

they put RISC-V in a soldering iron?

For anyone pining for innovation in Desktop, a small part of this culture is still alive in Ricing competitions.

A recent favorite of mine is this one. Timestamp starts at the final submission being reviewed: https://youtu.be/DxEKF0cuEzc?si=mqE_2vpKDBsMWlKW&t=557

Thanks for whoever preserved these! The CartoonNetwork website was one of my most fondest memories from my childhood.

These days the official website redirects to their YouTube channel which I feel is very sad. There used to be places for kids on the internet, now everything is heading towards major platforms which I honestly feel is going to be damaging the youth in the long term.

Upvoted for the username

Valid point. We have minimum age requirements set on some rules to avoid absorbing every latest change instantly.

Oops, my bad. We keep calling it Renovator internally but the name is RenovateBot or Renovate.

https://docs.renovatebot.com/

Who are their highly funded closed-source competitors they claim Warp cannot beat on price?

Warp is the only closed source terminal product I know of. Most other popular terminal emulators are open source already.

I feel like their funding is drying up and this is their last ditch effort to have the "community" build their product for them.

They claim agents will run the show, with inputs from community in the form of ideas/specs/direction. I wonder how long that will be sustainable for given the subsidized model prices are collapsing as we speak.

Is this an attempt to pivot to something else while the "community" keeps their first product alive? Maybe I'm being too cynical here, but I don't see this as an act of good faith, especially given their roots in VC funding.

I feel pretty happy we use Renovator (EDIT: It's Renovate) at my current workplace which by default will raise PRs to change any tags for actions with the SHA instead. Then, even when it bumps the version in future PRs, it bumps the SHA (with a comment of which tag version it represents)

Why does documentation require hosting it on a server? My assumption is that it's a static site, and as such, even GitHub Pages would be sufficient.

I know... all content has to be served via a "server" but in case of OVH it's a full-blown hosting solution isn't it?

Besides, I'm sure GitHub wouldn't mind supporting Pandas documentation. They do it for a million other projects for free (even though they're not popular among the HN crowd these days)

I decided to get back into reading two years ago and I picked this as one of the first ones to get started with, given it was a small book. I absolutely love Arthur C. Clarke's style of helping you visualize the grand scenes.

His books are more plot driven and the characters are pretty flat, but it's so damn fun to read through!

Morgan Freeman has been trying to get the movie adaptation made since early 2000s and wants to play Commander Norton. I had read that Denis Villenueve (the same director from the new Dune movies) was attached to direct the adaptation, but it seems like his schedule is really busy. He recently finished filming Dune Messiah and then he's got the next James Bond movie to deliver.

I seriously believe that it's not that GitHub is run on AI-generated code that's responsible for these slew of outages recently. I think it's crumbling under the load of a significantly large amount of AI-enabled coding with users raising PRs and pushing content a lot more than previously.

Obviously, if this is true, the team at GitHub is failing to scale their infra to meet the workload demands.

I place considerable doubt on claims of LLMs improving the user's thought process.

Especially since everyone harps on about it but never provides concrete evidence. If your thinking has sharpened, surely you can find a way to demonstrate how.

I suspect it's one of those things where the user thinks they have improved but the reality is different.

There's a research paper from the University of Liverpool, published in 2006 where researchers asked people to draw bicycles from memory and how people overestimate their understanding of basic things. It was a very fun and short read.

It's called "The science of cycology: Failures to understand how everyday objects work" by Rebecca Lawson.

https://link.springer.com/content/pdf/10.3758/bf03195929.pdf

If that's the goal, the technology for how these agents "learn" would be the most interesting one, even more than the demos in the link.

LLMs can barely remember the coding style I keep asking it to stick to despite numerous prompts, stuffing that guideline into my (whatever is the newest flavour of product-specific markdown file). They keep expanding the context window to work around that problem.

If they have something for long-term learning and growth that can help AI agents, they should be leveraging it for competitive advantage.

Servers, Ready to Eat

You only notice this stuff if you use shell very often, and practically live in the command line. Since I started my career I've been using omz and a fresh install is always snappy but over time it starts getting slow.

Debugging/profiling why it's gotten slow has mostly been an uphill battle for me. I tried using zprof which pointed that compdef and compinit were culprits. I tried changing my config to calculate compinit only once a day since most people reported it to work, but it never worked. This kind of stuff pokes and stabs at you endlessly.

OMZ being shell, and being a maze of a codebase, I couldn't track down if and where compinit was being called from even after the config change above, because all profilers pointed to the possibility of compinit being called twice.

I gave up and started using barebones zsh + starship because I do need a good prompt. Yet the issues persisted.

I recently started using fish + starship on my local machine so that I could evaluate it before committing to it at work. It's the fastest shell so far (maybe because its new, I intend to find out).

My only painpoint now is I have a bunch of utility functions I've maintained in bash that I need to port to fish because of the posix incompatibility.

OMZ is absolutely bloat.

We keep a precomputed cityhash64 value for a few columns we know are going to be used for aggregations. Rather than relying on ClickHouse to do it internally, this explicity behavior I've found is faster.

Especially if it's a multi tenant architecture, it helps to have the cityHash64 caclulated as a combination of tenant ID and another column, so the overall amount of data scanned is lowered too.

You're referring to the on-prem Jira. That might suck, sure. My experience has been purely using Jira Cloud and Confluence Cloud, both of which I've found to be snappy and responsive.

I find this to be a very amusing critique. In my experience, Notion (when I stopped using it 3 years ago) was slow as molasses. Slow to load, slow to update. In comparison, at work, I almost exclusively favor Confluence Cloud. It's very responsive for me.

We have tons of Confluence wikis, updated frequently.

If you're getting such value out of LLMs, I'm intrigued to learn more about what exactly it is that you're feeding them.

People boast about the gains with LLMs all the damn time and I'm sceptical of it all unless I see their inputs.

https://kanishk.io I just blog about a few things now and then.

I once ended up on the frontpage because of something I wrote: https://news.ycombinator.com/item?id=41689159

Dope!