semi

you might be interested in https://github.com/poettering/diskomator

tldr: bootable image to expose all disks over nvme-tcp.

Note that authentication and encryption is not yet implemented so I would be cautious, but it's still pretty interesting

you could even train an llm on examples of ai generated code you've edited to look like your own and use that to rewrite ai written code

You already specify 'ai programs'(plural) so what's the benefit of having a known port for unknown unquantifiable number of vaguely AI related programs?

who gets to decide what use of energy is providing sufficient value?

Just sticking to GPUs.. you see PoW block chains and LLMs as a waste, but is it any less wasteful to use the same GPUs at the same power draw to run a game at uncapped fps? or at whatever absurd resolution it takes to max it out?

Id rather we just charge for the energy usage, probably with some kind of tiering, and let people decide if their output is worth it.

who gets to decide what use of energy is providing sufficient value?

Just sticking to GPUs.. you see PoW block chains and LLMs as a waste, but is it any less wasteful to use the same GPUs at the same power draw to run a game at uncapped fps? or at whatever absurd resolution it takes to max it out?

Id rather we just charge for the energy usage, probably with some kind of tiering, and let people decide if their output is worth it.

if you want to get fancy and/or over engineered you would use systemd templated units to setup sshd@.service and a ListenAddress in the config listening on %i. Then you could bring up sshd@(expectedip).service for each expected IP

.. but that doesn't gain all that much tbh.if anything the only hesitation I'd have on listening to * and relying on firewall rules is if the service comes up before its configured. but exposing sshd isn't even that bad

I think that's more two independent byte streams. You want low latency but what is transfered doesnt really impact the other side, you just constantly want to push the next frame

why would a gui app be more readable for reading text? The only inate difference I can think of is the ability to vary fonts and rendering based on what you are rendering more.

And that can be a nice improvement to e. g show your code at a normal size but pop a tooltip up with smaller text, or have the linter/errors tab be a smaller font.

But that's a difference and not necessarily an improvement, because having a consistent font and fixed width text can make things more predictable and faster to interact with as you don't have to scan around as much.

disabling agent forwarding is the important bit.

But if you do want to break up your keys more, make sure you specify IdentityFile and Identities Only in the per host definitions in your ssh config.

By default assuming you use an ssh agent (no forwarding) with multiple keys and a default ssh config, the behavior is to just try to auth with every key in order.

So if you're worried about the ssh server identifying you, you're still exposing yourself. I don't think this is much of a concern but worth noting.

Slightly more important: you're wasting time during the initial connection to fail authentication a few times. This can matter more with higher latency

Even more important: sshd has a configurable number of times a client is allowed to fail authentication in a session attempt. If you have too many other keys in your agent you will just fail to auth before it tries the key that is actually valid for that host.

it depends on what you're needing suid for but the first thing id evaluate is if you can just grant a specific CAPAB instead.

> Also, if the ISP at home can do 10Gbps, 1Gbps, 300 Mbps whatever... I want to be able to use them with a single path, so there is no gain using multiple paths. Eventually, when I have cable+wifi connected at the same time, I use to force one of both, cannot see a reason to prefer using both at the same time. >

I don't understand why you would want to be able to use them with a single path. the gain would be being able to aggregate them and have individual tcp streams faster than any one IP connection could handle.

Though personally I think the resilience is more appealing. Not having to have a hard cutover when wifi degrades as I walk away would be nice

The other punishment for playing that way is how much you run into inventory and/or weight limits.

And yet it's still hard to convince myself to just skip looting all the useless junk

sounds similar to a bug in my predator x27 monitor where the middle strip of pixels moves to the far right side until I power cycle it.

The good news is it's just a firmware issue and they've already fixed it.. the bad news is to get the update they want me to ship a heavy ultra wide monitor to a service center.

so id check if your monitor has firmware updates and more importantly if you can apply them yourself.. and ensure any future monitor you buy actually allows users to apply updates to the firmware.

even online game servers can be questionable- like starcraft2 and diablo3(and 4) requiring them when they would be perfectly fine to play on lan only without any online servers like diablo2 and sc1

Or the many FPS games, basically all of them, that now require online servers ran by the company rather than letting users run their own the way all fps games used to. (and a dishonorable mention to the subset of CoD games that let you 'run your own server' but only by paying them to spin up another game server in their existing data centers)

or just claim ignorance. or the opposite and claim browser DNS prefetching when a blog spam comment linked to pornhub

It really depends on a lot imo.

If you still get days off it can be nice to have them during most peoples work week so you can get errands done easily. As long as it's not consistently every weekend so you can still do weekend things with other people.

It can also be easier to get stuff done on the weekend when most other people/distractions aren't around.

I actually am really happy with how my current team works, where you will be on call during the weekend but only expected to respond to pages so usually you just need to be able to respond and not actually working.. if lots of things page it's even more incentive to improve your systems to not get bugged on the weekend. But even if nothing pages you we still take two days of our choice off the next week.

I don't think you can fry a modern chip, at least without really going out of your way to.

But even if you just forced all of your fans to run at 0% (or physically removed them) I would expect the CPU to trigger PROCHOT and throttle it's performance, hard crashing if it has to.

>Ride sharing apps fundamentally sidestep this because you see the surge pricing without incurring a cost

Somewhat, but if your plan was to ride share home your choice might just be incur the cost or gamble that waiting will make it better and not worse.

That's always been the problem.. it's hard to justify the expense of making a game few people have the hardware to run.

So instead most games just target whatever modern consoles are which aside from right after a new release is usually relatively old for a PC.. which then makes it hard for PC gamers to justify spending that $1000+ when few games utilize it.

I haven't done it myself so I'm mostly speculating but I would expect a separate GPU to be easier to cool passively if it is performing as poorly as an iGPU would

At the end of the day it's all watts and surface area isn't it? Dedicated gpus are intended to perform better and allowed to draw more watts to do so. If you lowered the clock speeds and power limits down to what igpus are allowed to use, you'd then have a much larger surface to dissipate heat from and not share it with the heat from a CPU, so it should be easier to cool.

But id also expect anyone wanting to passively cool their system to also care about things like cost, size constraints, weight, number of parts that can fail etc so in practice just using an iGPU will probably make more sense.

> A court has no special access to facts

There are facts out there that can otherwise be kept private until subpoenad by court order, is that not special access to facts?

>. The key is just to not be too good; cheaters would shoot up out of my potential matches very quickly I think.

I've found the opposite to be true, at least to an extent.

Especially in the old counterstrike 1.6 days where 'ranked' was just privately organized matches from third party leagues, cheaters were much more of a problem for the lower ranked games because that's when you were playing against new accounts and there was no reputation on the line. Anyone who cheated would just make a new account and be right back in the entry leagues

Get good enough to rank up and you get to avoid most of the cheaters. some people still cheated, but it was also easier to spot when there are less players and you generally knew the players you play against.

These days I just play overwatch where I think both things are true. The automated matchmaking does let cheaters rank up a bit more but new accounts are still more likely to be cheating than established ones so getting to a higher rank let's you avoid a lot of them that would get banned before they can rank up enough

You can still do it off of one IP but it becomes more expensive as you now have to understand the protocol, and to route based on the handshake.

For tls1.2, that's usually routing based on server name identifier in the ClientHello and you can do it pretty trivially

For Minecraft.. you would probably need to write your own proxying logic. But as long as the protocol includes some info on what it's trying to connect to you could still do it.

Alternatively you could do it like the old quakeworld spectator proxies and just write a custom Minecraft server people connect to that then presents a menu in game to pick a server that it then connects to.

> Often we're talking about avoiding one db query per request in an application that is nowhere near any database-imposed bottlenecks.

The reason you avoid the db query isn't about trying to keep the db from being overloaded, it's that you can handle the request far away from the DB without needing to wait for the latency of hitting the DB.

Which is also still a complete premature and unnecessary optimization for most people. And then someone realizes their use case requires revocation so they add a DB check anyways.

that sounds like a dangerous scenario. If your docs are intentionally internal and not public, why would you let a publicly accessible LLMs answer questions with info from them?

An LLM trained on public docs for the public could be a better interface for projects with lots of public documentation.

An LLM trained on internal docs only accessible to internal users might be similarly useful

Even a private LLM on public docs for your support agents to use could increase their efficiency.

But I would never expose an LLM to the public that has been trained on data I don't want public

I could be wrong but I think it's only saving power if pixels are actually solid black where the pixel can be turned off entirely. I don't think dark shades are more power efficient than other colors. Thankfully a lot of apps are getting explicitly named OLED dark modes to take advantage of that.

it'd be nice if they picked them in alphabetical order

that seems like a chicken and the egg problem though. If you knew jqs syntax well enough to not need to use chatgpt, you might use it much more often.

in the hot ones reveal for this he mentioned using pepper x distillate in the sauce. isn't that just another word for extract or am I misunderstanding?

this is also how self extracting executables work. It's just a program that extracts whatever zip file exists after the program itself. Rename it from exe to zip and you have a valid zip file that just happens to have windows executable code before the start of the zip contents