Domains you could register an email with them at.

Ugh, yeah.

They were started by 4chan users, back when that sort of thing was more accepted there as “edgy humor”. Glad they grew up.

Yeah, I don’t know why they don’t have the normal “what is this” text from their main page at least at the bottom or something.

cock.li offers free email service with no personal info needed for signup. They’ve done this for 11 years, with no major outages to my knowledge, relying only on donations without explicitly asking for them or bugging their users.

I think at some point they also offered paid VPS services.

Very useful for accounts that you don’t want connected to your other “identities”, but where you’ll still need them associated with a real email for things like password resets.

It’s also tor friendly.

Flashing your lights or highbeams to pass is seen as quite agressive in the US. It isn’t taught in drivers ed, and the general interpretation I hear most people have of it isn’t “Hey, could you let me pass?” but instead “Hey! Fuckface! Stop driving so goddamn slow and get out of the damn way you shithead!”

Doesn’t help that in my experience, the only people flashing to pass are aggressively tailgating me when I’m already 10 mph or more over the speed limit.

Better to just pass on the right if there’s room. And if there isn’t room, fuck off telling someone to get out of your way. Not like they can get over anyway.

Some supposedly do, and then have been found hallucinating non-existent ones.

Guess the Snowflake (third party data storage and processing company/service) breaches earlier this year taught us nothing.

For additional context, this was not a choice, but a requirement. The Linux Foundation is US based, and Torvalds is a US citizen. This was required due to current US sanctions against Russia, and was not just some sort of “Russia bad” thing from Torvalds that a lot of people are framing it as.

Ugh. Righteous ideas about how things should work don’t change the fact that these network appliances doing it the wrong way still have years of time left before the bean counters consider them depreciated and let us replace them. Or that we’re locked into a multi-year contract with this business system that requires updating certs through a web UI.

Yes, there are almost always workarounds and ways to still automate it in the end, but then it’s a matter of effort vs stability vs time savings.

I love automating manual sysadmin actions, it’s my primary role on my team. Still, ignoring the complications that will unavoidably arise in trying automating this for every unique setup is incredibly foolish.

Yep, remote start.

They previously had remote start by key fob. They now only offer remote start via phone app with a $10 a month subscription.

This should realistically be part of every company’s disaster recovery/business continuity plan.

With NAT existing, I’m not sure there’s a significant reason to switch anymore.

Plus the “surprise” privacy and security benefits of just… not having every network connected device directly addressable by anyone else on the global network. The face of the internet and networking in general, plus the security and safety concerns around it, have changed dramatically since v6 was first created.

How do you pronounce github? GIMP? GNU? GPU? Javascript?

Oh Geremy, it’s time to jo to the jocery store! We need some jrape gelly.

So for those not familar with machine learning, which was the practical business use case for “AI” before LLMs took the world by storm, that is what they are describing as reinforcement learning. Both are valid terms for it.

It’s how you can make an AI that plays Mario Kart. You establish goals that grant points, stuff to avoid that loses points, and what actions it can take each “step”. Then you give it the first frame of a Mario Kart race, have it try literally every input it can put in that frame, then evaluate the change in points that results. You branch out from that collection of “frame 2s” and do the same thing again and again, checking more and more possible future states.

At some point you use certain rules to eliminate certain branches on this tree of potential future states, like discarding branches where it’s driving backwards. That way you can start opptimizing towards the options at any given time that get the most points im the end. Keep the amount of options being evaluated to an amount you can push through your hardware.

Eventually you try enough things enough times that you can pretty consistently use the data you gathered to make the best choice on any given frame.

The jank comes from how the points are configured. Like AI for a delivery robot could prioritize jumping off balconies if it prioritizes speed over self preservation.

Some of these pitfalls are easy to create rules around for training. Others are far more subtle and difficult to work around.

Some people in the video game TAS community (custom building a frame by frame list of the inputs needed to beat a game as fast as possible, human limits be damned) are already using this in limited capacities to automate testing approaches to particularly challenging sections of gameplay.

So it ends up coming down to complexity. Making an AI to play Pacman is relatively simple. There are only 4 options every step, the direction the joystick is held. So you have 4ⁿ states to keep track of, where n is the number of steps forward you want to look.

Trying to do that with language, and arguing that you can get reliable results with any kind of consistency, is blowing smoke. They can’t even clearly state what outcomes they are optimizing for with their “reward” function. God only knows what edge cases they’ve overlooked.

My complete out of my ass guess is that they did some analysis on response to previous gpt output, tried to distinguish between positive and negative responses (or at least distinguish against responses indicating that it was incorrect). They then used that as some sort of positive/negative points heuristic.

People have been speculating for a while that you could do that, crank up the “randomness”, have it generate multiple responses behind the scenes and then pit those “pre-responses” against each other and use that criteria to choose the best option of the “pre-responses”. They could even A/B test the responses over multiple users, and use the user responses as further “positive/negative points” reinforcement to feed back into it in a giant loop.

Again, completely pulled from my ass. Take with a boulder of salt.

Sorry buddy, but that capitalistic framework is where we all have to exist for the forseeable future.

Giving corporations more power is not going to help us end that.

Expecting any part of the brain to work that simply is foolish. We already know that REM sleep does a lot of the lifting for forming long term memories and processing complex input from your waking hours. Not to mention the importance of it for actual rest/recharging so you aren’t an exhausted zombie all the time.

Sounds like a wonderful idea to just fuck around with that.

Wow, that headline is significantly overstated. This has nothing to do with “understanding” or “reality”. Like holy shit this is an absurd leap from what they actually state in the abstract. Thanks OP for including it in the “preview” here.

Given a number of “starting states” and “ending states” of 2D grid based “environments”, an LLM is able to return increasing accurate “between states” as it is trained.

Neat! Not an understanding of reality. At the biggest stretch it is an “understanding” of an incredibly constrained problem space.

If it’s a concern for a corp, they can just ship the devices through the mail if there’s less scrutiny through the mail system (there usually is, they aren’t going to check the contents of every drive or try to crack the pin on every phone), or they just store their data in the cloud and pay for a new device for use at the destination so it never crosses the border.

… in the next 5 years?

Yep, and it’s harder to fix Windows VMs in Azure that are effected because you can’t boot them into safe mode the same way you can with a physical machine.

Your computer was likely not powered on during the time window between the fucked update pushing out and when they stopped pushing it out.