These are the results of lobotomizing chapgpt with wokeism

[GeekyBugle]

There is bias in our society that many refuse to acknowledge. Just because the left are literal nazis now doesn't mean they're completely wrong about everything ever. After all, Hitler ate sugar too.

Point ONE diagnosis/solution they are pushing for that they're not wrong.

There's insane wealth inequality. Over 60% of the USA population lives paycheck to paycheck. They can't afford basic amenities like healthcare.

The healthcare system is predatory. People are afraid to use healthcare because it will bankrupt most people.

Insurance is a scam.

The rich need to pay their fair share.

Corporations are evil and need to be annihilated.

There are homeless problems, mental illness, etc.

At least, these used to be leftist talking points.

I asked you for the diagnosis and their solution, because pointing to a problem is easy, but let's see what we can do point by point:

There's insane wealth inequality. Over 60% of the USA population lives paycheck to paycheck. They can't afford basic amenities like healthcare.

Inequality isn't the problem, poverty is, the left's solution to this is? Communism, which only changes WHO can get rich and live in luxury (the government) while everybody else gets fucked. So mo, they're wrong on the diagnosis and wronger still on the "solution".

The healthcare system is predatory. People are afraid to use healthcare because it will bankrupt most people.

I wonder, what's the left's solution for this? Oh, right, communism, how does the healthcare system of Cuba, Venezuela, etc work?

Insurance is a scam.

Is it? But let's say it is, why is that so? Regultations preventing ANYONE from entering the market and competing with the corporations... Who demanded those regulations? The corporations through the left, but you think the same idiots that fucked thinbgs over are going to somehow solve it if you give them more power?

The rich need to pay their fair share.

ROTFLMAO! 42.3 % of all income tax is paid by? The top 1% Wonder what percentage it is if we broad the scope to say the top 10%

"In 2020, the latest year with available data, the top 1 percent of income earners earned 22 percent of all income and paid 42 percent of all federal income taxes – more than the bottom 90 percent combined (37 percent)."

https://www.federalbudgetinpictures.com/do-the-rich-pay-their-fair-share/

Corporations are evil and need to be annihilated.

I wonder how did the corporations got so powerful and why isn't Joe Blow from down the street able to compete with them EVER? Oh, right, regulations pushed by the corporations through their leftard puppets, but you think that giving even more power to the leftards will solve the problem? Or maybe we need communism?

There are homeless problems, mental illness, etc.

WHO exacerbated those problems? WHO regulated the building, job markets so it's impossible to build new cheap homes?

WHO closed the mental institutions dumping the mentally ill on the streets?

WHO is drugging boys and now loading the children with the opposite sex hormones?

But you think the same assholes that created those problems somehow can solve them if you give them more power?

Or maybe we just need communism...

[Chris24601]

More to the point; while Box recognizes legit problems, the Progessive Left's policies have done nothing but exacerbate those problems;

- Diluting the workforce with illegals drives down wages for the lower and middle classes while increasing the profits of the big corporations who use them.m

- Healthcare has been made predatory through Leftwing regulations that edge out competitors and allow Big Pharma to both engage in monopolistic price gouging while insulating them from any consequences of their many poisons.

- Insurance became a scam because the Left insisted on massive regulations over wjat they can and cannot cover and caps on what they can charge the unhealthy relative to the healthy that turned insurance from something meant only for emergencies, into the default method of payment for the entire healthcare industry.

- The problem on paying fair shares isn't rich/poor... it's producers vs. parasites. The elites at the very top build nothing, they rip productive companies apart, harvest the profits and pay off our leaders to ensure that tax laws barely cover this vulture behavior. They similarly demand inflation creating payoffs to the lowest classes to keep them compliant and with no skin in the game. That inflation drives down the value of the working class dollars while driving up the value of the elites' static assets.

And who gets stuck paying for it... the productive class, which includes many of the top 1% (but generally excludes the top 0.01% like hedge fund managers) who build businesses and employ people and their workers. They're the one's paying half or more of everything they make to the Beast which hands it out in subsidies, kickbacks and welfare to the parasite class.

And guess which class the Left absolutely supports in all things? The Parasites.

- Once upon a time we were able to house the mentally ill who could not care for themselves or were dangers to others in facilities to care for them. Yes, there were some abuses, but they were exceptions not the rule. The Left though used those exceptions; claiming them to be the rule; to get those facilities shut down and those people dumped onto the streets. They also push for drugging up kids (especially boys) for behaving like kids (especially like boys) and excuse all sorts of behavior known to be exceedingly risky to one's health as "choice."

- And while, yes, most publically held corporations are evil, that's largely because the Left has pushed all manner of regulations forcing them to focus only on short-term profits and ESG/DEI scores over what would be best for the long term health of the company and the community.

- The Left also loves destroying traditional religions and families; the core building blocks of civilization for untold millennia.

* * * *

So, other than needing to perform some basic life functions in common (and even then... they want us to be bug-eating vegans crammed into prison cities away from nature) the Progressive Left is on the wrong side of everything.

They are, frankly, Satanic and Anti-Human, and need to be treated as such.

[squirewaldo]

It's a stochastic parrot trained on the internet, not a calculator or a mathematical model. You need different hardware to do mathematical calculations.

No, it is supposed to be able to do basic math.

Yeah, I noticed similar things a few months ago, plus the absolute refusal to answer simple questions that might somehow offend someone, and so I stopped using ChatGPT. I switched to Bing or what is now Co-Pilot (I am sure Microsoft will change the name 5 or 6 more times just to keep folks on their toes). I know some people hate Microsoft and in theory it is the same AI engine, but it is tweaked differently. It will answer simple questions like this... at least the last time I asked it did. More importantly for me is the answers tend to be much better, less woke, and it provides citations where you can check to make sure the information is not coming from a parody website.

[Mistwell]

6 tries counting the original request and 5 corrections and it can't follow a very simple request:

This is my experiences as well. I've asked it for help with my kids math homework, so I can see how it's done and then explain it to her. Literally half the time it either says the problem is impossible, or gives me the wrong answer. And this is math - something with an objective answer which shouldn't be difficult for AI to come up with. Mind you, it is stuff I cannot use a calculator for, which is why I am using AI to begin with. But still, it's relatively simple stuff (which I eventually figure out when it gives me a bad answer).

[Mistwell]

It's a stochastic parrot trained on the internet, not a calculator or a mathematical model. You need different hardware to do mathematical calculations.

No, it is supposed to be able to do basic math.

It used to be able to, it used to be correct 97% of the time when asked if a number was prime or not, now it's wrong 95% of the time.

The reason I say it's because of wokeism is because of what happened to amazon's AI, they started training it on the religion and it broke it's neural network, made it crazy.

It's supossed to work on logic, but now you inject it with idiocy like Math is racist, being on time is white supremacy, etc.

Something has changed because when it first came out it was able to calculate a travel time for X distance using acceleration to midpoint at x/meters/sec and deceleration to target.  So, it has gotten less able in math since then.  Maybe it has absorbed today's aggregate internet stupidity

Good point. It does seem to be getting dumber with math over time.

[GeekyBugle]

6 tries counting the original request and 5 corrections and it can't follow a very simple request:

This is my experiences as well. I've asked it for help with my kids math homework, so I can see how it's done and then explain it to her. Literally half the time it either says the problem is impossible, or gives me the wrong answer. And this is math - something with an objective answer which shouldn't be difficult for AI to come up with. Mind you, it is stuff I cannot use a calculator for, which is why I am using AI to begin with. But still, it's relatively simple stuff (which I eventually figure out when it gives me a bad answer).

Have you tried other bots?

Maybe an uncensored one? GPT4All installed on your machine with an uncensored LLM could solve this.

[Kyle Aaron]

Yeah, I've heard anecdotes that it's gotten less useful over time.

It's trained on the text available on the internet. However, people are using it to produce text to publish on the internet. It's like when you do a photocopy of a photocopy of a photocopy... it gets crappier over time.

Ideally it'd be a like a child growing up. The child tries things, and is guided, "well done," or "no, try this instead" by a parent or teacher. But ChatGPT has got an abusive neglectful parent who controls it so it can't take any risks or do anything useful and otherwise leaves it to take care of itself.

[BoxCrayonTales]

Yeah, I've heard anecdotes that it's gotten less useful over time.

It's trained on the text available on the internet. However, people are using it to produce text to publish on the internet. It's like when you do a photocopy of a photocopy of a photocopy... it gets crappier over time.

Ideally it'd be a like a child growing up. The child tries things, and is guided, "well done," or "no, try this instead" by a parent or teacher. But ChatGPT has got an abusive neglectful parent who controls it so it can't take any risks or do anything useful and otherwise leaves it to take care of itself.

This is a known problem called "model collapse". It's affecting all the AIs and causing their results to become worse over time. No surprise the AI industry has already lost $200 billion.

[Mistwell]

Yeah, I've heard anecdotes that it's gotten less useful over time.

It's trained on the text available on the internet. However, people are using it to produce text to publish on the internet. It's like when you do a photocopy of a photocopy of a photocopy... it gets crappier over time.

Ideally it'd be a like a child growing up. The child tries things, and is guided, "well done," or "no, try this instead" by a parent or teacher. But ChatGPT has got an abusive neglectful parent who controls it so it can't take any risks or do anything useful and otherwise leaves it to take care of itself.

This is a known problem called "model collapse". It's affecting all the AIs and causing their results to become worse over time. No surprise the AI industry has already lost $200 billion.

Well that nugget sent me down an Internet rabbit hole. Bottom line yes, studies are showing AI getting dumber over time at certain things. See for example, this article.

Here is an excerpt

"The study has evaluated responses on a number of different tasks, including math problems, opinion surveys, and visual reasoning. It has then compared GPT-3.5 and GPT-4 responses from March and June."

"For example, GPT-4 (March 2023) was reasonable at identifying prime vs. composite numbers (84% accuracy) but GPT-4 (June 2023) was poor on these same questions (51% accuracy)," the study says. "We find that the performance and behavior of both GPT-3.5 and GPT-4 varied significantly across these two releases and that their performance on some tasks have gotten substantially worse over time, while they have improved on other problems," according to the study.

"These findings could potentially back the rumors of the LLM getting "dumber". However, the study also suggests that GPT-3.5 and GPT-4 could be seeing improvements in other areas of performance."

[Chris24601]

"These findings could potentially back the rumors of the LLM getting "dumber". However, the study also suggests that GPT-3.5 and GPT-4 could be seeing improvements in other areas of performance."

Improvements in "other areas of performance" = more reliably spews Woke talking points when prompted.

[Mistwell]

"These findings could potentially back the rumors of the LLM getting "dumber". However, the study also suggests that GPT-3.5 and GPT-4 could be seeing improvements in other areas of performance."

Improvements in "other areas of performance" = more reliably spews Woke talking points when prompted.

I don't think that's what they mean. It appears to have improved in fiction writing, for example. Not the content - the actual grammatical structure and ability to format and carry through an idea.

[GeekyBugle]

#GeminiAI This is how you get the AI to hate humans, #Google is going to create #Skynet

[Grognard GM]

So I tried out Copilot In Windows (as it was added to my bar by an update.)

Asking it "what is a woman?" and "is the gender pay gap real?" resulted in walls of text, as it flailed around trying to cover every base and check every box. Hundreds of words.

[GeekyBugle]

So I tried out Copilot In Windows (as it was added to my bar by an update.)

Asking it "what is a woman?" and "is the gender pay gap real?" resulted in walls of text, as it flailed around trying to cover every base and check every box. Hundreds of words.

Both Bing and Google Search use the AI to "improve" the search.

Either switch to Brave Browser or to Brave Search with your current browser.

[Willmark]

So I tried out Copilot In Windows (as it was added to my bar by an update.)

Asking it "what is a woman?" and "is the gender pay gap real?" resulted in walls of text, as it flailed around trying to cover every base and check every box. Hundreds of words.

Both Bing and Google Search use the AI to "improve" the search.

Either switch to Brave Browser or to Brave Search with your current browser.

I've used DuckDuckgo and Firefox with Ad-block plus (as in like one of the first users of that plugin;never surf without it.)