I heard he made a line of surfing products, starting with Chump Wax.

The ol’ choke and stroke

11 was when I learned to masturbate. 12 was when I first heard the term “autoerotic asphyxiation”. I very well could have been a statistic…

One string pulling up the left side, one string pulling up the right side. They are separated in the “down” position, so they have to be separated in the “up” as well.

If you use only one string in the middle, they will never stay level.

If you want something done right, you’ve got to do it yourself.

If you already know what is wrong and just need a doctor’s note (and maybe antibiotics), go to the clinic. While their staff are significantly more skilled knowledgeable than the general public, their policies limit them to only simple diagnostics and treatments. Your medical knowledge is certainly less than that of the Nurse Practitioners and Physician Assistants that staff these clinics, but likely exceeds the scope of practice they are limited to by their employer. If you don’t know what the problem is, the clinic is going to refer you to your PCP or urgent care anyway, so you should only visit the clinic to appease HR or get access to basic prescription medications.

If something is bothering you, but you can tolerate it for a couple weeks, schedule an appointment with primary care.

If you don’t know what’s wrong, or you need something more than a note and a prescription, and you can transport yourself, go to urgent care.

The only time you should go to the ER voluntarily is if urgent care sends you there. Any other trip to the ER should be because someone dragged you there without giving you a choice.

I’m glad you mentioned insulin pumps, because there is a community of developers working on pumps, making them available to a broader audience, providing more people with better control over their blood sugar levels than manufacturers are willing or able to provide on their own.

https://openaps.org/

What you are arguing for is a threat to systems like OpenAPS, and to the people who benefit from them.

People repair their brakes wrong all the time. It’s absolutely caused accidents.

It also allows end users to install parts superior to OEM, improving braking capabilities, and preventing accidents.

Any automotive technician can tell you that manufacturers take engineering shortcuts, resulting in a product with certain deficiencies. The manufacturer’s motivation is to put out a product that widely appeals to the general public. They want nothing to do with a product specifically tailored to the needs of a particular individual.

We probably shouldn’t let people repair their own brake pads

What kind of auth-dystopian nonsense is that?

Repair an insulin pump the wrong way and it will absolutely kill you

You’re just as dead if you can’t get that insulin pump repaired or replaced because the manufacturer won’t or can’t support it. When they go bankrupt because other customers have sued them into non-existence, you still own the device they manufactured, and you still need it repaired.

Further, you presume the manufacturer can provide the best repairs. It is entirely possible and plausible that a competing engineer or programmer can improve upon the device, rendering it safer or providing superior operation. Car Mechanics can install a better braking system than the cheap, generic calipers and pads provided by the factory. Repair technicians can replace generic parts of medical devices allowing superior operation.

Yes, dangers exist from third party repairs.

Refusal or even simple failure to provide critical repair data to the end user or their agent denies the end user the ability to make an informed decision about repairs.

The company should be liable for all damages from a botched 3rd-party repair unless they provide to the end user complete specifications and unrestricted access to the device in order to make informed decisions about repairs.

Proprietary information and corporate classified information do not exist once they are incorporated into the device and sold to the end user. That information now belongs to the end user, who will continue to need it even if the company is out of business, or refuses service to the owner of the device.

Any attempt to conceal that information from the end user should make the company liable for any failed repair performed by any individual, including harm arising from that failed repair. The only way to avoid that liability is to release all information to the end user, so they are fully informed when making a repair decision.

I need this exact pin. You can save some money and make it without any moving parts.

You may need gasoline, but you don’t need BP’s gasoline. By choosing to buy BP’s gasoline, you support everything BP has ever done. Don’t want to support them, buy different gasoline.

FWIW, I’m not sure if I have a Xitter account or not. I did at one point. Definitely don’t remember a password, and I probably used a former email account that I can no longer access either, so no way of recovering it if it still exists. I have a severe lack of fucks to give about it.

But, I am pro-pedantry, and your argument kinda sucked.

The “collapse” you’re talking about is a reduction in the diversity of the output, which is exactly what we should expect when we impart a bias toward obviously correct answers, and away from obviously incorrect answers.

Further, that criticism is based on closed-loop feedback, where the LLM is training itself only on it’s own outputs.

I’m talking about open-loop, where it is also evaluating the responses from the other party.

Further, the studies whence such criticism comes are based primarily on image generation AIs, not LLMs. Image generation is highly subjective; there is no definitively “right” or “wrong” output, just whether it appeals to the specific observer. An image generator would need to tailor itself to that specific observer.

LLM sessions deal with far more objective content.

A functional definition of insanity is doing the same thing over and over and expecting different results. The inability to consider it’s previous interactions denies it the ability to learn from it’s previous behavior. The idea that AIs must not be allowed to train on their own data is functionally insane.

Also, with llms there is no “next time” it’s a completely static model.

It’s only a completely static model if it is not allowed to use it’s own interactions as training data. If it is allowed to use the data acquired from those interactions, it stops being a static model.

Kids do learn elementary arithmetic by rote memorization. Number theory doesn’t actually develop significantly until somewhere around 3rd to 5th grade, and even then, we don’t place a lot of value on it at that time. We are taught to memorize the multiplication table, for example, because the efficiency of simply knowing that table is far more computationally valuable than the ability to reproduce it at any given time. That rote memorization is mimicry: the child is simply spitting out a previously learned response.

Remember: LLMs are currently toddlers. They are toddlers with excellent grammar, but they are toddlers.

Remember also that simple mimicry is an incredibly powerful problem solving method.

I can see why you would think that, but to see how it actually goes with a human, look at the interaction between a parent and child, or a teacher and student.

“Johnny, what’s 2+2?”

“5?”

“No, Johnny, try again.”

“Oh, it’s 4.”

Turning Johnny into an LLM,nThe next time someone asks, he might not remember 4, but he does remember that “5” consistently gets him a “that’s wrong” response. So does “3”.

But the only way he knows 5 and 3 gets a negative reaction is by training on his own data, learning from his own mistakes.

He becomes a better and better mimic, which gets him up to about a 5th grade level of intelligence instead of a toddler.

What other networks?

It currently recognizes when it is told it is wrong: it is told to apologize to it’s conversation partner and to provide a different response. It doesn’t need another network to tell it right from wrong. It needs access to the previous sessions where humans gave it that information.

It needs to be retrained on the responses it receives from it’s conversation partner. It’s previous output provides context for it’s partner’s responses.

It recognizes when it is told that it is wrong. It is fed data that certain outputs often invite “you’re wrong” feedback from it’s partners, and it is instructed to minimize such feedback.

It is not (yet) developing true intelligence. It is simply learning to bias it’s responses in such a way that it’s audience doesn’t immediately call it a liar.