That’s true, the problem with the original statement is that it is too broadly scoped by “knowledge”, implying that it is any and all knowledge. If I obtain the knowledge to write a singleton in object oriented programming while at work - even if the concept is applied to a work project, and later use the programming concept of a singleton in my own software, then they can’t do shit.

A simpler example that shows that it’s too broadly scoped is that if I get trained and certified to use a forklift for a job, and later start my own company and have to use a forklift, there is no precedent for my original employer to come after me for using a forklift in my business operation just because I learned how to use a forklift while I worked for them.

If the knowledge is proprietary or copyrighted or a trade secret and what I do uses any of that, or what I produce is a 1 to 1 product of that, then they can come after me.

Oh ok, no worries.

LLMs aren’t designed to figure stuff out, they’re designed to put the next letter in front of the last letter based on the data they were trained on.

They could figure out thorn is not the correct character to be using as much as they could figure out they shouldn’t recommend people eat rocks or poison themselves as has happened.

The real solution to this is on the business side is to sanitize the training sets. Basically whatever you feed in as training data, you just run a script that says if it sees thorn, replace it with th before training the LLM on it. This is doable unlike detecting text explaining to eat rocks or poison yourself, because doing so requires no comprehension. For thorn it’s just a find and replace operation.

This is kind of how VeraCrypts hidden partition feature works.

You start the process of the volume’s encryption and set a “false” password for it. It creates a partition that is encrypted with that password. When it finishes, you mount it and store “fake” files, the files you would reveal under duress. Veracrypt then takes in a second password and creates a “hidden partition” in the remaining free space of the disk - to be clear, that memory space still reports as unused/free if investigated, but the partition is there.

You can then mount that with your second password and store your actual files. You can work with files and folders in the hidden partition as needed, however if anything is added or changed etc in that first fake partition, the data in the hidden partition will be corrupted by those actions.

This means that so long as you plan ahead, someone can literally put a gun to your head and demand the password to the encrypted disk, and you can give them one that works without revealing the data to them.

In theory, since the data in the hidden partition is encrypted and unreadable, it is impossible to detect that it exists in the “unused” space of the disk, even by a forensic analyst. To them it would just look like old, randomly flipped bits that came from previous usage followed by a quick format.

Now, what’s really cool about this is that if you use the veracrypt bootloader, you can store and boot from an undetectable OS you store in that hidden partition, while having a decoy operating system on the visible partition:

https://veracrypt.io/en/VeraCrypt Hidden Operating System.html

I got so good with lucid dreaming when I was younger that I could pretty cleanly drop straight through full consciousness into the half sleep stage of paralysis and then into unconsciousness in which I would immediately find myself lucid in a dream, but having been aware of everything in between.

This is quite difficult to do from what I recall because you have to completely clear your conscious mind of all thought while maintaining awareness of what you are intending to accomplish in some lower part of your brain. It’s quite hard because you have to be aware but not thinking. Practicing meditation is about this exact kind of thing, and while I have never really done meditation, I expect practicing it would help a lot for this.

I could actually feel the point at which I went into the paralysis and had no input to my body anymore, and the oddities of perception beginning to shift due to my half awake brain.

When I actually dropped through that floor into sleep (which for me did not take that long after the paralysis from what I could sense), the feeling of moving through into actual sleep was pretty wild to experience as it is something usually not remembered or noticed, and I have no great way of describing how it was for me other than it kind of felt like I folded or collapsed inward on myself.

Sort of like if everything around you and every sense you had rushed away suddenly. My sense of sight, touch, smell, taste, hearing, balance - but not like it all just disappeared instantly, more like it was all “pushed” off of me over the course of a moment or two, which is why it’s just such a difficult and strange thing to try to describe. In a way it did feel kind of like falling - in the least a sense of motion is the closest thing that could be used to describe it, but that also is not really it either.

Then I was immediately aware of the dream and knew that I had accomplished the goal, after which I was free to release the strong awareness I had been maintaining and turn my focus to what I was actually looking to do in the dream otherwise.

So yes, you can learn it and get quite good at it, but I think it depends on the person, a lot of practice, and a lot of trial and error. There are other methods that can be used to discover the dream mid-way through as well described in this thread, but you can enter the dream at the beginning in my experience too.

Oh for sure - I think that this method has more efficacy in production environments ran by small businesses anyway, since best practices are rarely followed in many of them (until something happens that changes their mind on what they budget for haha), and even at that it is still a rare attack to see.

I am unaware of this type of attack ever occurring on a persons personal network, most likely because so few end users make backups, there is no need to go through the trouble of doing this, making this method useful only in highly targeted attacks.

We are definitely in agreement on proper backups still being the best method to recover from the vast majority of problems - even this one, depending on the backup solution.

They usually embed themselves in within the system files and have some scheduled job that basically checks for the criteria - if you are only backing up and restoring user data then it’s a non-issue, but if you do a full recovery including the system files/the system scheduler etc, then it can happen, and it is often necessary to backup executable and system files for production environments (true, not so much for individual users and their systems).

When I was working in an IT shop, one of our clients was ransomwared with this method. The saving grace for us in that instance is that our backups were going to a product that allowed you to easily break open and dissect the compressed backups pre-recovery, so we were able to determine where the malicious files were and kill them before pushing the backups. Of course we only noticed that it was in the backups after we had tried to push the backups once already, so it was quite the timely process - I think I worked for something like 18 hours that day.

You can read about such malware if you search for “timebomb malware” or “malware does not execute until date” etc.

The attack is not super common anymore, but still happens.

For example, here is an article discussing time bomb methods on linkedin.

https://www.linkedin.com/pulse/time-bombs-malware-delayed-execution-any-run

Another on the knowbe4 blog:

https://blog.knowbe4.com/ransomware-can-destroy-backups-in-four-ways

There are viruses that are time-bombs. They specifically don’t do really do anything until some criteria is met in the future, such as the current date being beyond a specific date, at which point they proc. They do this in order to make sure they are in your backups when you restore them so that they immediately run when recovery is completed and the system is booted.

little one deleted

Before replying, I will note that I feel as though we have reached the end of the discussion - I think you have raised your contentions well, and I understand them but disagree. I expect you feel the same from your end, and that is fine, but I am sure neither of us thinks we will really convince the other of anything at this juncture.

As a result, I won’t be continuing the conversation after this, but know that I don’t feel any animosity toward you.

Yes, and there have been cases of guns not going off and failing to kill anyone but that is a very pedantic take… The fact of the matter remains, guns are designed to kill people, other things could kill people but not been designed for such purpose, they tend to be less effective

Some guns are designed to kill people, others are designed to hunt, others are designed to target shoot.

For example, you don’t see one of these killing a lot of people:

https://www.ssusa.org/media/c0yk1ziu/12feinwerkbau-aw93.jpg

Note that an argument of “it could be used to kill someone however” returns us to placing it in the same category as a vehicle.

The point is that, since seemingly we all agree (even gun owners as per your comment)… why do we do it at all when we all agree it’s a bad idea?!

Poor regulation depending on your area. Recall that I am in agreement that regulation can always be better. I disagree with the regulations for this in a place such as America, but you should examine how this works in other countries as well (such as the paper I linked regarding Switzerland).

Not quite… I would not trust a toddler to get a pie out of the oven because, no matter how well trained, such toddler will likely burn themselves and ruin the pie. Sure, Larry is a disaster, but we have COUNTLESS examples of Police Officers, arguably the most trained demographic to hold guns, who constantly misuse them.

Yes and I would not trust a toddler to drive either, just like how I would not trust Larry to drive, nor to use a firearm. Officers (in the US I assume you are referencing) do not receive nearly as much training as I think you suppose. Furthermore this again goes back to regulation - I believe that if a cop misuses a firearm or a vehicle, then again, they should have those things taken and be jailed. This again hints that you are more upset with specific people/regulatory systems than firearms or vehicles I think.

You don’t seem to have produced an argument against guns that does not directly depend upon a specific group of people choosing to misuse one, but the same argument can be applied to my car bomb allegory.

The amount of people that could truly be trusted with guns, under special circumstances, is very very slim. No amount of education or training would make a human 100% trust worthy with guns 100% of the time. There is a reason a huge percentage of violent crime falls in the category of “passion” crimes

The same could be said for someone driving - most likely more often for drivers since more people own vehicles than those who own firearms. It is anecdotal to say that is slim as well - you should search for a source to back that up in the future. I suggest you look into the actual data regarding gun ownership versus responsibility.

I would highly recommend you read the paper I linked in my first comment regarding Switzerland because it implies the opposite in their case.

In the US? Yes that could be different, but again that demonstrates quite clearly this is not a gun problem, but a societal/educational problem since this problem only really exists in specific places.

We should do those things… and still not let almost anyone own a gun. The case is clear, there is simply no societal benefit to allow widespread ownership of certain guns.

I don’t think the case is very clear at all - based on the sources I have provided I would say it is decidedly unclear. Social benefits include the control of wildlife, military protections, and social sport (such as olympic shooting, and target shooting competition).

I used to feel similarly to yourself so I challenged my bias by going through the process of getting a firearms license in my country and engaging with firearms, sport shooting, and the community that surrounds it. After all, if my bias did not change, then I could simply sell the firearms to recoup my money and would not have lost anything - however it did change my opinions on the matter, so just know that I am speaking from a place of having involved myself with the subject directly to go seek out the real tangible information on my own behalf.

Perhaps someday you may find it enlightening to do something similar, even if you don’t take it quite as far as me.

Thanks and have a good day.

the harm people can do improvising an everyday device as a weapon, is magnitudes of order lower

Not necessarily.

https://en.wikipedia.org/wiki/2018_Toronto_van_attack

You could also fill the car with a lot of gasoline canisters and fertilizer if you so wished. These are all also a lot easier to get than a firearm, particularly if you are crazy.

Agree that regulation can always be better however.

How can anyone see that and think “oh sure Larry is a crazy person, drives drunk all the time, usually on his phone, but I am sure he will be super responsible with an automatic machine gun”

I think you are drawing a false causal relationship/strawman here. Almost no one thinks this, including 99% of people who own and use firearms. Certain people should be prevented from owning and operating firearms and certain people should also be prevented from owning and operating vehicles.

A person who operates a vehicle irresponsibly should have their license and vehicle taken and be jailed in such a case.

A person who uses a firearm irresponsibly should have their firearms/firearms license taken and be jailed in such a case.

Such a person using either thing irresponsibly can result in the loss of life, but I don’t see as many people trying to ban vehicles, gasoline, and fertilizer because they are capable of killing multiple people.

To me it sounds like the issue you have is not with vehicles or firearms, it is with Larry. This brings us back to my point about this being a societal/educational problem rather than a banning problem. I get the feeling if Larry wants to hurt a lot of people, he will find a way to do so regardless. If you want society to be safe from Larry, you would have to go a lot further than banning only firearms.

Alternatively, you tackle the societal responsibility/education/mental health problems that society has, and maybe Larry stops drinking, gets therapy for his mental problems, gets off social media and now feels as though there is no need to hurt anyone or to act irresponsibly with guns, vehicles, gasoline, knives, baseball bats, tire irons, or whatever else.

I believe weapons should be banned and that crime should not exist in the first place

A car can be used as a weapon as can cleaning products, baseball bats, tire irons, kitchen knives, sharp sticks… etc. If someone wants to purpose something as a weapon, then they will.

Crime is defined by law and law is defined by government and/or society. As long as people exist, crime will exist. It is not sound reasoning to believe “crime should not exist” because if it were made illegal to wear black shoes, crime exists again, and as such it is an impossible standard.

Rather, I accept that crime will always exist in the world as a result, but aspire to a world wherein there is no real need to ban things like guns because no one uses them to harm other people - the same goes for cars, baseball bats, etc.

Banning registered/licensed owners from owning firearms does not do much, because the last thing a potential mass shooter does when obtaining a firearm is register or get a license. As such, laws that ban only really affect people who are generally responsible in the first place.

If all firearms suddenly disappeared, people would just build rudimentary ones if they wanted one for violence as well. Shinzo Abe was killed by a gun someone built in their home. To prevent that you would have to make the purchase of metal piping and whatnot illegal as well.

Stopping mass shootings, gun violence, and violence in general is not a matter of banning something, it is a matter of education and societal responsibility. Read about the comparatively high gun ownership yet low shootings in Switzerland for example.

https://www.sciencedirect.com/science/article/pii/S1359178924000776

Just an aside, but fellas comes from fellows which comes from the old English feolaga which means “one who lays down money in a joint enterprise” making it gender neutral despite contemporary usage.

Perhaps try pastels?

The Debian docs were really useful for me in setting up my 3090 on Debian proper.

Since Mint is downstream, maybe they will help you.

https://wiki.debian.org/NvidiaGraphicsDrivers

This is beautiful.

I’ve been looking for exactly this for some time, really excited to try it out.

Reverse proxy/SOCKS5 works well in my experience.

I have a little computer on my network which runs my VPN - then on that computer I have ssh listening on a non-standard port that my VPN’s dyndns links up to a human readable hostname with a different port.

If I want to watch stuff off-network I just have to ssh -D to that hostname and port and then configure a browser to use the connection as a SOCKS5 proxy, then jellyfin and anything else I’m hosting works as if locally through that browser.

The ssh is key based as well, not password based - haven’t had any incidents in doing it this way.

It wasn’t for me on Debian 12/13. I just had to add the repo for the drivers and run 1 or 2 lines of bash and I’ve been good ever since with my 3090.

I’ve found real solutions to pretty much everything but this. For Fusion, I still just have to run it in a windows VM under Linux.

I’ve found a lot of success after biting the bullet and purchasing bitwig as my DAW.

Interesting, thanks for sharing that, I will have to keep an eye on it as well.