The most interesting part of this take is that JD Vance is very much in the big tech friends circle with Thiel and Musk. But I’m sure they can run some antitrust against their enemies at Google or whatever

I remember one internship in college, I realized that after 4 months of work, the result was 15k lines less code than when I started. I figured out new ways to structure the system so it was much easier to write and maintain, while actually adding features. That felt great

And yeah, there are many ways for it to happen. Ex. someone was shipping the tests with the code and decided to stop, debug symbols being removed, inlined dependencies being externalized, maybe a new version of a UI toolkit has extra icons built in

Efficiency can gently creep in. What blows my mind is that this is averaged out across so many packages at once. And sure, sometimes it goes up too, but nothing like Windows/OSX. It’s really cool that you can make a Linux that will fit into ~any space you want, whereas the min requirements for Win11 include 64gb of hd

Out of all the comments here hahaha this is the one that gets me lmao

You haven’t had an update with less than 52 packages?? Ever??

Well, I took the time to read the whitepaper, and it’s yeah, pretty dumb sounding. The gist is that it’s p2p post sharing with lots of captchas & a crypto edge that it probably doesn’t need https://img1.wsimg.com/blobby/go/eb02f20b-e787-4a02-b188-d0fcbc250ba1/pleb.tex-6d2e1bf.pdf

The similarities to Lemmy are substantial, it’s just not on activitypub, but rather its own pubsub thing. If you want to host data, you still have to keep a node running at all times, it’s not the case that “there are no instances”. Those instances can moderate the content, so it’s not the case that “there’s no moderation.” The whitepaper mentions that “its possible to delegate running a client to a centralized server…” rather than having to have a fat syncing client running on your own machine … in lemmy, it’s more like “its possible to run your own node if you want”. Plebbit doesn’t care about maintaining history of posts, it expects that servers will go down over time, and the data will be lost. Lemmy is pretty similar in that regard too, if all instances hosting the data go down, then it’s lost. The expected outcome is that there’s a handful of big nodes, as is the typical result of this form of “decentralization” - same as Lemmy, Email

Ultimately, I don’t see Plebbit doing anything particularly smarter/better, and having private/public key cryptography involved doesn’t really matter. They talk about blockchains and using coins as anti-spam mechanisms, but I don’t see why that’s relevant to the implementation

Then I read that chat apps and YouTube would not be banned, and scoffed

Literally chat apps are social media. You can post stories and pump memes and news. You can even have bots that scrape and post content. YouTube is just a matter of checking a box whether it’s “for kids” and they already do that. Basically the whole thing is stupid

Y’know, you’re right & that’s wild. I guess I should have known, but didn’t assume that they have like 600m in unrelated investments. Though the burn rate is quite a lot too, so they probably would scale back browser dev a lot if it lost its profitability & become a pure VC kinda org

Not only does it need to do everything from memory management to job scheduling, it also has all of the UI and graphics driver complexity blended in. Usually that’s a different layer that the kernel historically didn’t worry about, it would be as if GTK is part of Linux, along with the programming language. Then there’s shit like WebAssembly and WebGL, databases, sandboxing, permissions, user management… A Brower is like a cross platform OS built to run on another OS

The thing is it’s never been more expensive and time consuming to write a browser, it’s bigger scope than a kernel in many ways. Stuff like Epiphany isn’t even close, despite relying on Apple’s webkit. Most distros just push people to Firefox now, despite a history of KHTML and all that. We would need something like the Linux Foundation to pick it up (which runs on corporate sponsorship for a shared resource)

Yeah but in the short term the company will literally go out of business

And somehow breaking even more sites than safari; missing many modern features

Welllll I mean technically webkit is totally separate from chromes blink fork for the last several years. I.e. browers like safari and gnome’s epiphany. That said, safari is mac/iOS only and epiphany sucks ass so…

Nah he was saying he was okay with free versions of his app undercutting him before, but calling his paid version a scam caused him to reconsider the policy - threatens revenue

That’s like one thing ML can actually help with XD cute cat

Truly a golden era

I’m down but if it can’t handle a 5s dc then rip

I just got a new laptop and was genuinely gonna try windows 11 and wsl for my coding needs. But in first boot, it demands internet to do updates. Ok, I connect to coffee shop wifi. Nope, won’t do it because it can’t handle the click through screen to accept wifi ToS. Fine. I take it home, where my Internet is great but has a glitch where it drops out for a few seconds now and then. Turns out that windows will literally cancel updating and demand I reconnect and restart for the kind of drop that I barely notice day to day. So I gave up, plugged in my ArchLinux thumb drive, and mkfs.ext4 before rsyncing my entire old computer to it

I’m kinda annoyed that this whole thing was pretty much a pitch for Tauri, and that’s a pretty lame looking webapp thing with typescript and whatever browser engine you happen to have lying around

Tauri is tryna be all like “hey look at our install size, it’s smaller than electron!!” … like anyone cares about install size much. The problem is the memory/cpu use of web apps, which tends to 5x a decent native app. Maybe one day, with webassembly…

Facepalm again and again every time my non technical boss asks me if Ive been using genai to speed up my work. No boss, I haven’t, that actually slows me down

You can collect the data and figure out how to use it later. Just look at the Google leaks lately and what they collect, it’s literally everything down to the length of clicks and full walks through the site

Collecting data about user interests is in itself valuable, and it’s plausible to use various metrics to analyze it, something as simple as sentiment analysis, which has been broadly done. Sentiment analysis has predated modern ML by a long margin, but you can read the wiki page on that

But yeah just think about stuff like Google trends, tracking interest in topics, as an example of what such data could be used for. And deanonymizing the inputs is probably possible to some degree, aside from the obvious trust we place in DDG as a centralized failure point

These companies absolutely collect the prompt data and user session behavior. Who knows what kinda analytics they can use it for at any time in the future, even if it’s just assessing how happy the user was with the answers based on response. But having it detached from your person is good. Unless they can identify you based on metrics like time of day, speech patterns, etc