the main reason why sched_ext is taking off is that EEVDF and CFS before it are -general- cpu schedulers, so while they work well for most tasks, they don't work as well as they could in certain situations like heavy server loads, low-latency loads (like multimedia production and gaming), or security sensitive loads. These schedulers are built and or specifically tunable to these types of loads, such as scx_lavd, scx_bpfland, and scx_cosmos. They each approach things differently, such as using core compaction, sharing or unsharing per core task queues, being way more cache context aware, keeping hot cores hot, etc. And the perf boost in these specific contexts is very real. Heck, Meta started using scx_lavd for better task load balencing over EEVDF on their servers.
There are plenty of workplaces that require a more secure development environment. Centralizing the dev environment without having to go full PC over IP would be a very cost friendly way to manage this.
Nyxt is a new interesting browser. WebKit engine but configured by lisp under the hood, and fully keyboard controllable with emacs, vim, or cua controls.
nice..had an idea of a browser engine powered by a different scripting language like Lisp. Even better, a browser engine that can plug in any language runtime through which web apps can be coded (WebAssembly is working in that direction to provide a single portable runtime I think).
I think Microsoft is smarter than that, look at Github for example, people have largely forgotten it's owned by MSFT. They would have left Discord as a gaming platform and used it's tech in Teams instead. Teams is so behind Discord in all areas except video quality.
Very poorly polished stuff, to be honest. Before, they always seemed to take great care in making their core product (and the API for it) great. The only thing I can think of that was similarly unpolished was GitHub Enterprise.
I don't think GitHub staff has forgotten. Actions are straight up Azure Pipelines. The code is a huge mess. If I didn't know better I'd say it's obscured by design. Check out the actions/runner repo if you don't believe me. They polished them very well for the hosted version, but the cracks show if you try to use the self-hosted version.
I can't believe anyone at GitHub is particularly thrilled about having a Microsoft technology that broken imposed on them.
Can I just say, while I believe Discord will die anyway if it IPOs (as with anything that believes in infinite scaling), I would have loved to see Microsoft enact the special fucked up kind of integration that only they can manage
This is how tech startups work, isn't it? Use investors to build a product with no business model, then the founders cash out via purchase or IPO, and soon after the fact that you have no business model comes to light and the product falls apart trying to find a business model, while users move on to the next shiny thing with no business model.
Glad I'm not the only one thinking this. Seems obvious, but I always felt like I was missing something. How is it that most of the tech/SV world is just the same con being run over and over again, yet it keeps working? Does it operate on the same psychological glitch as a lottery? Do the giant corps buying these startups all think "I know how it's been in the past, but surely THIS startup WON'T turn to shit the moment I buy it because it was never a real business to begin with! This time for sure!"
Or are the real suckers the investors, and the corps are just doing these purchases in a performative capacity to keep razzle-dazzling them?
A lot of actual profits have been made by mature tech companies. Valuations seem too high right now, but I don’t think it’s a giant fraud. More like there’s nothing better to invest in.
Yes, the model is entirely built on acquisitions, where discord itself can't be profitable, but part of Microsoft it can deliver value by deepening the mote around everything else.
Many things are only valuable as a public good or part of monopoly. Such is funny relationship between monopolization and socialism.
Discord has a business model: Get people engaged in a community, sell them Nitro so they can boost their communities [1][2]. The "buy benefits for you community" scheme is wildly successful in mobile games, so I wouldn't be surprised if it works well for Discord.
The fact that they are trying to sell/IPO is pretty strong evidence that they are not profitable, I think. But regardless, I will give them massive props for not just doing the ad thing, which is the last gasp of this style of startup before they finish circling the drain. Lookin' at you, Imgur.
I would in all honesty give them even bigger props, as someone who used to be but isn't a fan anymore. They tried to add a game store a la steam into it but realized it didn't get the engagement they'd hoped for so they stopped putting time and effort into it and shut it down.
My first impression when they first added it was that they'd just shove it down people's throats and keep trying to make it work.
I hope I'm not wrong about them getting rid of it again and I've just gotten used to tuning all that out when using discord.
This is what flailing around trying to find a business model looks like. If Imgur was profitable without those ads, they wouldn't be abusing their users like this.
The second they try to monetize discord like a mobile game is the second all of my friends stop using it and hop to the next silicon Valley chat app that pops up
I’d slightly revise your description of the scheme from no business model to a intentional no revenue business model allowing for the valuation speculation to run rampant (i.e. at anytime we could stop investing in our growth, then it’s all profits). Of course by the time they go public like say Uber and set multiple records on quarterly losses and the investors drop the bag on the public it becomes obvious they can’t stop spending or the business will go under so instead they will continue accumulating billion dollar plus losses every quarter so by the time the shit hits the fan original founders and investors are on to the next thing and can always say they took a xx billion dollar startup public and that’s when the company lost its culture and the corporate greed ruined it.
Corporate greed is for sure a problem, but... does Discord even make money? I'm guessing their S-1 will reveal that they are ludicrously unprofitable, and they will actually have to fix that if they go public, corporate greed or no.
In general, I try not to fall in love with free products that seem too good to be true, because the party has to come to an end at some point.
At least on my hardware Skype and Teams are much more responsive than Discord though. And the laptop's coolers get the workout of their lifetime whenever Discord spins up.
They'd probably turn this IM into family friendly safe bay with strictly moderated content - because let's be honest, nowadays anything goes in. Discord accounts at first would be offered an optional merge with MSA and in long time, you'd have to log into using MSA credentials only. Microsoft Discord branding would arrive. There would be a business oriented version created and MS would abandon Teams; basic Discord would have ads related to your activity - you could avoid that (along with telemetry) by purchasing a subscription as the current monetization options would be removed. A special version with github/git related features would be created - for free, but only for those who are really using it in code related tasks.
Yes, but Microsoft would probably prefer to offer such subscription among own plans - like 356, and that would also happen after Discord would get Microsoft branding. It's a pure guess of course but I think it could happen.
Deeper Xbox integration, probably. Microsoft knows the writing is on the wall that Xbox Party Chat usage is way down and a lot of it has been replaced by gamers with Discord.
They know it so well the new Xbox Wireless Headset essentially has a "Discord feature" though obviously not named or marketed directly as that. (It supports dual pairing with a bluetooth phone for "phone calls over game audio", but everyone I know got the message between lines that it supports Discord chat, and were talking about the headset precisely because of that feature.)
Microsoft would have bought Discord for the gaming community and pursued tight integration with the XBox ecosystem. Which could have boosted the platform, but probably not in a way that resulted in a lot of direct revenue.
The S-1 will reveal all, a tight integration with Xbox and not a lot of direct revenue may still ultimately result in a sustainable and viable product, once you go public indefinite quarterly losses don’t get saved by the next funding round at higher valuations.
its more like Vivaldi's tab stacks or FF's Tree Tabs. With groups there are no inherent additional security mechanisms in place to prevent cross-group contamination/access. i.e., two different tab groups can access the same site data.
containers however are completely isolated from each other. i.e., two different containers have completely different sets of site data.
> no inherent additional security mechanisms in place to prevent cross-group contamination/access
Are you saying the biggest advertiser on the planet likely doesn't want to add functionality that could potentially hinder tracking for advertising purposes...?
Elaine Chao’s sister is married to Xi, while Elaine, as transportation secretary under Trump, was busted inviting family with business ties to the CCP to official US government meetings.
The fear on this forum is imagined political thriller more than realistic.
Every technologist is grifting off the military industrial complex.
I still don't understand what they mean by "loss". Why is this even a problem? The information isn't gone, its right -there- in the black hole, which until proven otherwise, is part of the universe.
Until someone can prove the universe cares whether the info is in a black hole or not, its not really a problem is it? If anything the universe usually shows it doesn't care what we humans think, its going to do its own thing, regardless: i.e., weak nuclear force and "symmetries"
> The information isn't gone, its right -there- in the black hole
No, it isn't; it hits the singularity inside of the hole and gets destroyed. At least, that's what Hawking's original model, the one he used to predict that black holes evaporate, says.
One way of seeing why Hawking's model had to say this is to combine the following facts about the evaporating black hole and the Hawking radiation in Hawking's model:
(1) The hole itself cannot contain any information other than its mass, charge, and spin (because of the "black holes have no hair" theorem), which is far too little information to describe everything that fell into the hole.
(2) The Hawking radiation cannot contain any information about what fell into the hole because it is thermal, black-body radiation, i.e., the only information it contains is its temperature, which is related to the mass of the hole.
So the information can't be stored either inside the hole or outside the hole, which means it must be destroyed, and the only place it can be destroyed is by hitting the singularity inside the hole.
The black hole information loss problem is that the above is inconsistent with quantum unitarity. So Hawking's original model can't be right; but nobody knows what model should replace it.
> Maybe the information gets encoded in digits of value of mass expressed in some unit.
No, it can't, not all the information. Two objects of the same mass but different internal composition would add the same mass to the hole, but would be described by different information. So the hole can't store in the value of its mass which of the two objects fell in.
More generally, a hole of, say, ten Solar masses could have gotten that mass by an infinite number of possible combinations of things falling in. The mass itself can't distinguish between any of those possibilities; all it can tell you is that ten Solar masses total of stuff fell in.
But why do you assume two objects of same mass but different composition will add the same mass to the black hole? Different composition means different interactions during the fall and different amount of radiated energy. Infinite number of bits can be encoded in single real number. It is hard to measure more than few digits of it, but so it is hard to measure all the information.
> why do you assume two objects of same mass but different composition will add the same mass to the black hole?
Because that's what the physics says. See below.
> Different composition means different interactions during the fall and different amount of radiated energy.
All of that can be taken into account before the object falls into the hole; the observer outside can measure it all and deduct it from the mass he expects to be added to the hole.
We are talking about the mass that gets added after all that; and for any given mass added to the hole after all those things are taken into account, there are many different possible combinations of objects falling into the hole that can add that mass.
> Infinite number of bits can be encoded in single real number.
We are not talking about math, we are talking about physics. The number of bits that can be stored in an object of finite size is finite as far as physics is concerned.
> "for any given mass added to the hole after all those things are taken into account, there are many different possible combinations of objects falling into the hole that can add that mass."
Approximately, sure, best scales can do around 5 significant digits and null measurements can get us few more digits. But we can't verify equality of mass to arbitrary precision. For elementary particles of same kind, we can assume their masses are the same. But there is infinity of digits available. Perhaps there are no two differently composed bodies that have the same real number as mass (too many options to be different). Then maybe any mass addition to mass of the black hole can encode all the information there is about the body.
This is irrelevant to the argument; our finite ability to measure masses is not what we are talking about. We are talking about what masses are physically possible, whether or not we can measure all of them with unbounded accuracy.
> there is infinity of digits available
You can't have it both ways. If it is physically true that there are an infinite number of digits available to specify an object's mass, then it is also physically true that there are multiple possible combinations of objects whose masses can sum to that same mass (in fact there will be an infinite number of them).
Conversely, if it is not physically true that there are multiple possible combinations of objects whose masses can sum to a given mass, there cannot be an infinite number of digits available to specify an object's mass: there must be only a finite number of possible masses, and the numbers specifying the possible masses must be such that no two such numbers add up to another such number.
It is relevant because the limitation to our measurement capability means we can't know most of the digits. We can't confirm experimentally that a given mass can be composed of "multiple possible combinations of objects whose masses can sum to that same mass". The mass number for an object can exist and yet it may be impossible to duplicate it with other objects. Imagine every object having unique ID with infinity of digits.
> Imagine every object having unique ID with infinity of digits.
And then, as I said, there will be an infinite number of possible combinations of other masses that will add to that mass. The fact that we can't verify that experimentally is irrelevant; your model allows it and that means that, in your model, the unique ID of a given black hole's mass would not uniquely identify the original pieces of matter that formed it, and therefore would not provide the information that you originally claimed it could provide, in the post of yours that started this subthread.
Imagine a 2t object falls into the hole and then a 3t object. Can that be differentiated than what would have happened had there been only one 5t object using mass alone?
Only if mass conservation is broken, and current theory does not predict this (where does the extra mass go to?). Same applies for the other 'no-hair' theorem properties - spin and charge.
Sure, and the point of this video is that while that may be _mathematically and theoretically_ sound, there's no way you can realistically make any measurements or any observations to confirm or deny your particular idea. What we have a lot of these ideas, with no way to discern between theories which accurately represent nature and theories which are merely mathematically correct.
The black hole is not eternal. Once it is fully evaporated, you still need to account for the information that was contained within (or accept information loss).
I believe this was addressed in the first few paragraphs of the article: this problem is not about 'information', which is a vague phrase. Rather, it is about how black hole evaporation is fundamentally time irreversible.
Right, but that is using a semi-classical calculation, whereas we know that ultimately any process (evolution of a closed system like the universe) compatible with quantum mechanics needs to be unitary/reversible.
That mismatch is what sets up the BH information “paradox”.
All of Quantum Mechanics respect unitary evolution, unitary evolution can be rewinded back. Black hole evaporation breaks unitary evolution, at least in the semi-classical approximation of Hawking. If you prepare a pure quantum state it will come out as mangled thermal radiation. You cannot return to the pure quantum state from the thermal radiation (i.e. we have lost information about it).
This transition is impossible in Quantum Mechanics and it would suppose a killing blow to Quantum Mechanics if true. So a better way to rephrase our worries is that if Black Holes do not respect unitary evolution then our most precise physical theory is fundamentally wrong.
