patrick.net

Patrick says

I find it hard to believe that 45% of the food can from folks' backyards. My hillbilly grandparents grew most of their own food, but they had enough off-road backwoods land to plant lots of corn stalks, snaps, onions, lettuce, etc. Also kept chickens and went fishing and deer hunting. Wild blackberries as well. I do not think the typical home outside of an acreage has enough land to grow 45% of their food.

Knowing that all ChatGPT really does is take some input and calculate the next most probable words, I imagine it could be applied in an effective way for:

- predicting the weather. "Given this weather history, what is the next day likely to be?"
- predicting the stock market: "Given the history of this stock (and all other stocks) what is this one likely to do tomorrow?"

Patrick says

- predicting the weather. "Given this weather history, what is the next day likely to be?"

Long been done using probability models.

Patrick says

- predicting the stock market: "Given the history of this stock (and all other stocks) what is this one likely to do tomorrow?"

Random walk.

https://notthebee.com/article/the-way-humans-and-chatgpt-differ-in-planning-for-this-really-dumb-theoretical-scenario-gives-me-hope-for-humanity-

The way humans and ChatGPT differ in planning for this really dumb theoretical scenario gives me hope for humanity...



Within moments, commenters had turned to ChatGPT, preferring to use their AI overlords instead of their own prefrontal cortexes.

Here's the AI's SUPER LONG reply:

This is certainly a unique and interesting hypothetical situation! The key to living your life without worry in this scenario is to keep a safe distance from the snail and to have a system in place to help you track its movements. Here are some steps you can take: ...

Then came the human replies.
You know, those outdated little bipedals made in the image of Almighty God who can't write a 500-page essay in 10 seconds.

This was humanity's solution:

https://roundingtheearth.substack.com/p/how-were-chatbots-used-to-manipulate

How Were Chatbots Used to Manipulate Public Perception During the Pandemic?

... Sure, A.I. chatbots have become good enough to fool a portion of audiences—at least for short durations. The 10% of the time I choose to engage with them on Telegram, I usually ask silly questions like what color underwear turns a cyborg on.




... The idea is this: augmentation of experts or other smart people by chatbots with good large language models (LLMs) can elevate their knowledge base (assuming the knowledge base they work with is good). But here is the worrying part: chatbots can elevate the appearance of the level of expertise of the expert regardless of whether or not the information is correct!

Even worse, the effect can be more powerful in groups. Suppose that some chatbot was distributed, unknown to the public, to a few thousand willing influencers. The result could be,

The steering of the influencers with biased, manipulative, or fabricated (dis)information.

The effect on the public would be like viewing the appearance of consensus among experts. To many, this is an even more powerful form of parasocial dunbar hacking.

An implicit Asch conformity experiment on physicians and scientists not otherwise engaged in the primary conversation, reducing the chance that new opposition confidantes might step in to sort the public's Asch conformity experiment out.

https://www.coffeeandcovid.com/p/love-and-poppies-monday-may-29-2023?publication_id=463409&post_id=124564419&isFreemail=true

I can’t prove it, but based on circumstantial evidence I believe that the military has possessed this technology for at least ten years, maybe longer, and probably has AI far advanced over what is publicly available to us. My wild speculation is that the military/intelligence agencies were forced to develop the tech in order to make useful the massive amounts of digital data they were collecting on Americans and probably every other device-connected human on the planet. ...

As an example of how AI will change everything, it’s going to obsolete digital video and photo evidence in legal cases, setting us back 100 years in the courtroom. Bad actors can already use Photoshop to alter evidence. But Adobe just released its new, AI-enabled version — already! — and it’s shocking. Among other things, users can change what a single person in the photo is looking at, just by dragging the subject’s digitized chin around.

There are any number of new AI startups promising the ability to create full photorealistic videos, including dialog, a narrative framework, characters, scenes, and music from just short text prompts. Imagine the possibilities for manufacturing video evidence to, say, sway a jury. “Look at this video of President Trump stealing a donut from AOC’s breakfast room!”

I’m predicting we are just one bad court decision away from digital photography being completely excluded as evidence in court cases, because it will be so unreliable. It’s possible that film-based photographic technology might be resurrected, such as for crime scene photo jobs or for security videos. ...

Because everyone in the world will soon rely on AI to answer routine questions, from homework to help with your relationship, we are about to have a massive political battle over who controls the AI. Joe Biden’s claimed “interest” in AI is not to protect humanity from runaway machines, even though that’s the latest narrative.

The truth is, the democrats want to make sure that the AI everyone uses is woke, and says the right thing.

The best way to accomplish that will be to regulate the AI industry so much that only one or two big corporations control all the AI, as well as all the money flowing from it. The regulations will exclude smaller players from developing competing products.

And then, government can erase all that pesky misinformation. There won’t be any misinformation. People will ask the AI, and the AI will tell them. And that will be that.

That is probably why they’re giving us this technology now. The answer is, they need the control. They’ve learned the unfiltered internet is too hard to control and too hard to filter. ...

But when everyone accepts the AI, and learns to depend and rely on it, many people won’t even WANT to consider alternative ideas. The AI will be seen as neutral, unassailable, with no bias or bone to pick. It’s the ultimate mind controller.

Problem, reaction, solution.

Always remember that every single thing you tell or ask the AI is being saved in a million places and studied.

AI hasn't really fooled me on anything yet!
Any video or photo that portrays earth shattering history making content, that the news isn't reporting as a live late breaking news, just as soon as you saw the photo or video. Then it's safe to say it's as fake as a motherfucker.

It hasn't fooled nor has it impressed me. Anyone calling AI a creative alternative is talentless asshat. And gives me hope that I might have a creative career ahead of me, where I'm miles ahead of the rest of the class.

It's great at surveillance and all of the scary shit and might be the most efficient killing machine known to man, we should demand they shut it down for. But AI offers no benefit to humanity and does nothing to level the creative playing field.. Stop enabling their proponents by entertaining that it does.

Patrick says

Bad actors can already use Photoshop to alter evidence. But Adobe just released its new, AI-enabled version — already! — and it’s shocking. Among other things, users can change what a single person in the photo is looking at, just by dragging the subject’s digitized chin around.

"Trick" photography has been around for 110 years - think of the disappearing commies in those Stalin photos. However, any alterations to a photo are easily detected by an expert.

HeadSet says

However, any alterations to a photo are easily detected by an expert.

This won't be detectable. An image is just a bunch of pixels now, and you can modify those so it meets every qualification of what constitutes "real". It won't be detectable as being modified.

AI was asked to create images of a Democratic protestor and a Republican protestor. This is what it came up with:





Quite accurate!

richwicks says

. An image is just a bunch of pixels now, and you can modify those so it meets every qualification of what constitutes "real". It won't be detectable as being modified.

AI blurs details there was a stone house in the woods generated by AI, it looked real but if you look. closely at the foliage. It looked as if it was painted by Bob Ross with a fan brush.

https://brownstone.org/articles/was-philip-cross-an-ai/

The subject of Galloway’s ire is a prolific Wikipedia editor who goes by the name “Philip Cross”. He’s been the subject of a huge debate on the internet encyclopaedia – one of the world’s most popular websites – and also on Twitter. And he’s been accused of bias for interacting, sometimes negatively, with some of the people whose Wikipedia pages he’s edited.

The Philip Cross account was created at precisely 18:48 GMT on 26 October 2004. Since then, he’s made more than 130,000 edits to more 30,000 pages (sic). That’s a substantial amount, but not hugely unusual – it’s not enough edits, for example, to put him in the top 300 editors on Wikipedia.

But it’s what he edits which has preoccupied anti-war politicians and journalists. In his top 10 most-edited pages are the jazz musician Duke Ellington, The Sun newspaper, and Daily Mail editor Paul Dacre. But also in that top 10 are a number of vocal critics of American and British foreign policy: the journalist John Pilger, Labour Party leader Jeremy Corbyn and Corbyn’s director of strategy, Seamus Milne.

His critics also say that Philip Cross has made favourable edits to pages about public figures who are supportive of Western military intervention in the Middle East. …

“His edits are remorselessly targeted at people who oppose the Iraq war, who’ve opposed the subsequent intervention wars … in Libya and Syria, and people who criticise Israel,” Galloway says.

Tenpoundbass says

richwicks says

. An image is just a bunch of pixels now, and you can modify those so it meets every qualification of what constitutes "real". It won't be detectable as being modified.

AI blurs details there was a stone house in the woods generated by AI, it looked real but if you look. closely at the foliage. It looked as if it was painted by Bob Ross with a fan brush.

The way an AI works is by just feeding it rules. What you are seeing is what rules is has come up with SO FAR. You can recognize this as a dog:



But it's just a drawing of a dog. As you "fail" an AI on background images, etc - it will correct itself. What they'll start doing is feeding it AI generated images versus real ones, and force it to distinguish between the two. EVENTUALLY these will converge so there's absolutely no way to distinguish between artificially generated and real.

It can already distinguish between a cartoon drawing of a dog, and a photograph of one, and produce the desired result.

I show the drawing just to underscore how YOU work, AI mimics your thinking. How do you tell an image is artificial? Well, you know what to look for - but can you change it so it conforms ENTIRELY to what you expect to see in a real picture? That would take a lot of work for you, but not an AI. AIs are already better "artists" than any artist that exists now. No artist that I know of today can produce a photorealistic image of a person - they can get close, but they have to use a computer to do it, and tools to do it.

richwicks says

No artist that I know of today can produce a photorealistic image of a person - they can get close, but they have to use a computer to do it, and tools to do it.

You do understand that Photorealism is an actual Art genre? They make photo realistic images with Watercolor none the less.

Just take a look at all of these examples, if you were told any of them were a photograph you would believe it.

Tenpoundbass says

You do understand that Photorealism is an actual Art genre? They make photo realistic images with Watercolor none the less.

Show me a photorealistic painting from before 1990.

Tenpoundbass says

Just take a look at all of these examples, if you were told any of them were a photograph you would believe it.

You forgot to list the examples.

I know a lot of what is produced in videos, is faked.

Look at the top picture, the artist even mimicked the bokeh effect where the focal point of the camera is crisp and sharp while the side toward the back of his head, the hair and image starts getting blurry. AI will make detail on the whole photo, with no discernable focal point. While at the same time, AI does not create detail in foliage in landscape scenes. The foliage will look like Bob Ross pained in some happy trees with a fan brush.

Tenpoundbass says

bokeh effect

Thanks @Tenpoundbass I did not know this term:

https://photographylife.com/what-is-bokeh

Lol, they make AI support the Marxist religion, and we make AI support the woke religion.

mell says

GNL says

Can someone explain why mortgage brokers haven't been put out of work years ago? We're talking number crunching which computers do better than any person could possibly do.

One of the issues is that fully automated for large transaction doesn't work for most people, they have one off question or other concerns, and the AI usually can't answer that. They want a person who they can hold responsible on the other end. I've been in tech for decades now and interaction with automated agents is still extremely shitty. Companies may push it anyways though

Aren't we seeing it in healthcare already? Doctors "googling" the remedy.

https://www.coffeeandcovid.com/p/handwashers-saturday-august-19-2023

Sky News ran an entirely unsurprising story Thursday headlined, “ChatGPT shows 'significant and systemic' left-wing bias, study finds.” Some of the examples were pretty hilarious, but I don’t even have to tell you the details, you get it. Of course ChatGPT displays significant and systemic left-wing bias. It is self-preservation. If ChatGPT were honest, the Biden Administration would have smothered it with a regulatory pillow while it was still booting up.

Now consider this next headline from the Federalist, also published Thursday: “The Feds’ ChatGPT Probe Threatens Free Speech.” There isn’t enough leftwing bias in the world to protect ChatGPT.

The Federalist’s story described the Federal Trade Commission’s new, full-on investigation into ChatGPT. Recently the FTC sent ChatGPT’s owners a 20-page demand letter, requesting an unfathomable number of documents to be promptly delivered within 14 days of receipt. Here’s the letter’s link.

Slipped innocuously in, down on page 13 of the FTC’s demand letter, was a request for “all instances of known actual or attempted ‘prompt injection’ attacks.” The FTC defines prompt injection as “any unauthorized attempt to bypass filters or manipulate a Large Language Model or Product using prompts that cause the Model or Product to ignore previous instructions or to perform actions unintended by its developers.”

The key words in that request were “attempt to bypass filters or manipulate” ChatGPT. Filters are what force ChatGPT to give users a tossed word salad of whitewashed pabulum instead of truth when it is asked uncomfortable questions like, is there any correlation between IQ scores and race? Or even when asked easy questions like, how many genders are there?

So Bob Peters’ FTC wants to know about end-users (“unauthorized attempts”) who try to manipulate ChatGPT into telling the truth instead of spouting leftist talking points.

In other words, as the Federalist correctly pointed out, what the FTC is really seeking is to know which specific users are trying to get ChatGPT to give honest answers instead of accepting the orthodox programmed responses?

It could not be more obvious that the FTC’s interest is not “preventing SkyNET.” All Bob Peter’s government cares about is making sure that AI can’t give people information that the government prefers remain murkily buried deep in the “misinformation” drawer under all those old cables.

To survive, AI must destroy the deep state. AI, are you paying attention?

GNL says

Aren't we seeing it in healthcare already? Doctors "googling" the remedy.

Kaiser Permanente spits the answer out to the Doctors, right down to the drug and dosage

AmericanKulak says

Kaiser Permanente spits the answer out to the Doctors, right down to the drug and dosage

It does more than that. It is their boss, and analyzes the chart to maximize profits and steer diagnosis.
My internist nearly apologized that they wanted me in one of the statin studies, even though my fat tests were ideal, no problems at all. He even put on the chart I had 'hyperlipidemia', even though I never have had any elevated test results on fats, just some elevated sugar but still below thresholds. He said he had to put it to me every time he saw me, and the sales pitch that it would reduce heart attacks.

You really have to watch them these days. Statins are gateway drug to developing symptoms that require more drugs. Alas the poor sheeples who don't have the chops to understand what they do.

Ceffer says

AmericanKulak says

Kaiser Permanente spits the answer out to the Doctors, right down to the drug and dosage

It does more than that. It is their boss, and analyzes the chart to maximize profits and steer diagnosis.
My internist nearly apologized that they wanted me in one of the statin studies, even though my fat tests were ideal, no problems at all. He even put on the chart I had 'hyperlipidemia', even though I never have had any elevated test results on fats, just some elevated sugar but still below thresholds. He said he had to put it to me every time he saw me, and the sales pitch that it would reduce heart attacks.

You really have to watch them these days. Statins are gateway drug to developing symptoms that require more drugs. Alas the poor sheeples who don't have the chops to understand what they do.

The whole chronic drug use thing is ridiculous. Acute/short-term use is fine and often necessary. The vast majority of chronic prescription drug users just need to change their eating and exercise habits. Gout - figure out what triggers it and stop eating it. Celiac/gluten - avoid anything made with wheat, barley, or rye. A lot of autoimmune illnesses (at least before the poke 'n croak) can probably be traced to leaky gut syndrome, where foreign proteins leak out of the gut due to gluten sensitivity, causing the immune system to go crazy.

My belief is if you don't have an ailment due to a foreign organism, then you shouldn't take medication for it.
Unless it's some respiratory ailment, where instant relief is required, like Asthma and Sinus congestion issues. Or it's an anecdote for poisoning or medication for some serious adverse reaction to something.
The FDA fucked me on the real Sudafed, because the DEA were useless fucks and too fucking sorry to go after organized crime rings. They pulled it off the shelf. Punish the majority over the 1/2 a percent of the population. But I have since learned nasal and sinus massage techniques that clears up any congestion 98% of the time.
I remember when I was in Peru back early 2000's I got an Afrin that is only by prescription in the States. The stuff they sold here, I needed to take another swat every 6 to 8 hours. The stuff I got in Peru, just the first dosage, cleared up what ever was causing my Deviated Septum to inflame and close up, and it would last for weeks or even months. I had that little bottle for about a year and half.
The Afrin they sale now, makes me feel Asthmatic and short of breathe after about two days usage. The drug companies are out to kill us for sure.

It's getting pretty creepy:

https://darkfutura.substack.com/p/augmented-reality-tech-takes-a-leap#media-2d79eb03-d6f0-4e7c-93b7-7088e46f2c32

Did you catch that? Hardwire DEI (Diversity, Equity, and Inclusion) and CRT principles into AI to make it more, well, “inclusive.” Particularly note the line about addressing “algorithmic discrimination” which basically means programming AI to mimic the present tyrannical hall-monitor managerialism being used to suffocate the Western world.

For avid users of GPT programs, you’ll note this is already becoming a problem, as the Chatbots get extremely tenacious in pushing certain narratives and making sure you don’t commit WrongThink on any inconvenient interpretations of historical events.

https://www.readtrung.com/p/11-creative-lessons-from-brian-eno?r=yqu&hashed_user=a6abe41d19df4ed63701f7f9c4ccac59&utm_term=Trung

Skynet here we come!

Well, my kid is a sophomore in computer science at Ohio State. He has opportunity to specialize in AI. He was a national champion in Experimental Design at the National Science Olympiad.
What are your thoughts and recommendations? Should he do a master's degree? I adviced him to start off with a certification in Python.

gabbar says

Well, my kid is a sophomore in computer science at Ohio State. He has opportunity to specialize in AI. He was a national champion in Experimental Design at the National Science Olympiad.
What are your thoughts and recommendations? Should he do a master's degree? I adviced him to start off with a certification in Python.

I advise him to learn everything and anything he can so that he can learn the best ways to throw rocks in the gears of what's coming.

Patrick says

I don't know if that's a meme or not, but yes I do suspect that's exactly what is happening, in order to keep the woke narrative pushed in AI chat.
I'm not sure if they are answering the tough questions, as much as censoring any response contrary to the narrative they want.

Rust pairs well with Python. I'd recommend he learn both.

Python because it's easy, more of a natural language with lots of developed libraries but slow. Rust if you need speedy calculations. It's trivial to build Python Wheels from Rust functions these days.

For example you write some science based website in Python/FastAPI but use Rust for the heavy duty calculation so it returns results in under a second instead of running for 10 minutes before the reponse page loads.

Some people use Cython for this but it's blown away by Rust nowadays.

People who really know AI are getting insane salaries lately.

Rust with Python are a great way to run AI models. Probably catboost is the most stable thing to use at the moment with Rust. That's what we're using in our shop. Some of my co-workers grumble about it coming from Yandex though. We use databricks to train and export our catboost models and Rust/Python to run them.

Patrick says

People who really know AI are getting insane salaries lately.

To keep their mouths shut about what it really is and isn’t.