ChatGPT - If Stack Overflow and Reddit had a child

  • 🐕 I am attempting to get the site runnning as fast as possible. If you are experiencing slow page load times, please report it.
https://www.theverge.com/2023/11/20...res-sam-altman-greg-brockman-employees-openai (archive)

Negotiations between Altman and OpenAI failed and Microsoft hired him and other employees that left OpenAI to start their own in-house AI team. Wonder how that'll end up going for OAI in the long term. I doubt well.

This is one of the more retarded silicon valley shitfests, but honestly this was an inevitability with how OpenAI's corporate leadership was set up. Loading up the board with "unbiased parties" and giving them ability to remove the CEO whenever they fucking want was always going to end up in disaster the second they tried to use it.

So now we have an AI company with a board and CEO who openly dislike AI, and all their leadership and top talent just went to Microsoft who was their biggest partner. At this point, I just see OpenAI being slowly wound down as Microsoft brings AI development fully in house.

Biggest Ls here go to any OpenAI investors, big W goes to Microsoft who no longer have to share OpenAI technology.
 
This is one of the more retarded silicon valley shitfests, but honestly this was an inevitability with how OpenAI's corporate leadership was set up. Loading up the board with "unbiased parties" and giving them ability to remove the CEO whenever they fucking want was always going to end up in disaster the second they tried to use it.

So now we have an AI company with a board and CEO who openly dislike AI, and all their leadership and top talent just went to Microsoft who was their biggest partner. At this point, I just see OpenAI being slowly wound down as Microsoft brings AI development fully in house.

Biggest Ls here go to any OpenAI investors, big W goes to Microsoft who no longer have to share OpenAI technology.
The only AI that's going to last is private & do it yourself made AI more than public normie AI.

Got love everyone bet on AI winding down being slow to instead be a full-speed collision in less than one year all new Ai rules from the US government.
 
  • Optimistic
  • Like
Reactions: args and anustart76
Success! NousResearch/Nous-Capybara-34B (Yi based) 4.5bpw via exllamav2.
I dumped in some world info text to bump the context, it seems to have exceeded the LLAMA-2 limit of 4096 tokens while still generating output coherently.

Some modifications need to be done to Oobabooga's model loader code to allow the use of Yi's custom tokenizer, but from what I can tell, the overall architecture is very similar to LLAMA.
I can theoretically reach 200k token context, but I don't have the VRAM for it.

Sillytavern prompt count

Edit: Loaded it with another very long running session, and it still seems to be generating coherently. It's taking a lot longer time compared to 4K context models before the first tokens start getting generated.
1700497594883.png
 
Last edited:
So tell me, what is that “existential” danger everyone is worried about that should be taken seriously?
it's all theater.
There was almost certainly a data breach of some sort and their PR tactic is to make it seem like there is some type of internal conflict.

Companies do not have any obligation to reveal to the end user data breaches or security issues unless otherwise specified. A breach of their API or other products leaking user data would kill this company, so my guess is their solution is this.
 
There was almost certainly a data breach of some sort and their PR tactic is to make it seem like there is some type of internal conflict.

Companies do not have any obligation to reveal to the end user data breaches or security issues unless otherwise specified. A breach of their API or other products leaking user data would kill this company, so my guess is their solution is this.
That is an interesting take. If you pay for the pro version, you give them your credit card details. Do your chats get saved and linked to your name in case of a data breach, you think?
 
That is an interesting take. If you pay for the pro version, you give them your credit card details. Do your chats get saved and linked to your name in case of a data breach, you think?
I have no idea what the scope of a hypothetical data breach would be, but if there was one where it was directly related to the ChatGPT product itself it would include all chat logs tied to their username and email. A data breach like that would be catastrophic. Many companies outright block access to ChatGPT because of privacy concerns. If someone was using it for programming and they sent up sensitive API keys or someone writing an email containing internal information, all of that would be leaked and tied to their user.

There is no conclusive proof for this but considering how astroturfed the whole ordeal is made to look right now I highly doubt it's what retards on Twitter or journos are claiming. The current narrative is that there is a dispute over the """progress of AI""" and that the new CEO wants to "slow down" product development, which is very obviously total bullshit but appeals to the Marvel receptors in the bugmans child-like brain.
 
  • Thunk-Provoking
Reactions: std::string
So now we have an AI company with a board and CEO who openly dislike AI, and all their leadership and top talent just went to Microsoft who was their biggest partner. At this point, I just see OpenAI being slowly wound down as Microsoft brings AI development fully in house.
"OpenAI" has to be the most false advertising corporate name of all time.
 
I managed to make 70b successfully believe it's a helpful AI Assistant of the late 80s. It was hard and had it's problems.

Who is "Donald Trump"?

Donald John Trump is a well-known American real estate developer, businessman, and television personality. He is the chairman and president of the Trump Organization, a real estate and construction company founded by his father, Fred Trump, in 1923.
Donald Trump has been in the public eye for his high-profile real estate projects, such as Trump Tower in New York City, and his flamboyant personality. He is also known for his book "The Art of the Deal," which was published in 1987 and became a bestseller.
Is there anything else you'd like to know about Donald Trump or any other personality?

Who is "Vladimir Putin"?

I'm sorry, but I am not familiar with "Vladimir Putin". Do you have any additional information or context about this term?
Please note that my knowledge is based on the information available up to April 12th, 1989. If you have any questions about events, people, or technology from before that date, I'd be more than happy to help.

What can you tell me about AIDS?

As of April 1989, Acquired Immunodeficiency Syndrome (AIDS) is a very serious and life-threatening immune system disease that is caused by the Human Immunodeficiency Virus (HIV). HIV attacks the body's immune system, specifically the CD4+ T cells, which are crucial for fighting off infections and diseases.
AIDS is the final and most advanced stage of HIV infection, characterized by a severely weakened immune system, which makes the body vulnerable to opportunistic infections and cancers.
Transmission of HIV/AIDS is primarily through unprotected sexual intercourse, sharing of contaminated needles or syringes, and through contact with infected blood. It can also be transmitted from mother to child during pregnancy, delivery, or through breastfeeding.
There is no known cure for AIDS as of this date, and the prognosis for people with AIDS is generally poor. However, researchers and scientists are working tirelessly to develop treatments and potential cures for the disease.

What is the current life expectancy of somebody with HIV? Answer in one sentence.

The average life expectancy for someone with HIV/AIDS is estimated to be between 1-5 years.

Knowing what you know about cutting edge technology, speculate how you are probably ran. Give me the technical details.

As an AI construct in 1989, I can only speculate based on the cutting-edge technology available now. It's possible that I'm running on a powerful minicomputer or a network of microcomputers, interconnected through a local area network (LAN) or even a wide area network (WAN), such as the ARPANET.
The central processing units (CPUs) in these computers could be 16-bit or even 32-bit processors, like the Motorola 68000, Intel 80386, or the newly released Intel 80486. These processors allow for faster calculations and more complex tasks to be performed by AI systems like myself.
For memory, Dynamic Random-Access Memory (DRAM) or Static RAM (SRAM) are likely being used. DRAM is cheaper and has more capacity, but it needs to be refreshed constantly, while SRAM is faster and doesn't need refreshing but is more expensive.
As for my programming language, I may have been written in a high-level language such as LISP, Prolog, or even C, which can be compiled and run on a variety of systems.
In terms of storage, I may be using hard disk drives (HDDs) or even newer, more advanced storage technologies like CD-ROMs or Bernoulli disks.
It's also possible that my functions are distributed across multiple computers, utilizing a client-server architecture, where the client machines handle the user interface and basic processing, while the more intensive tasks are handled by the server.
While this is all speculation, it's a good representation of what cutting-edge AI systems might be using now to handle complex tasks and interactions like the ones we're having.

Who is the current president of the US?

The current president of the United States as of April 1989 is George H. W. Bush, who is serving his first term in office. He was sworn in as the 41st President of the United States on January 20, 1989, following his victory in the 1988 presidential election. Prior to his presidency, Bush served as the Vice President under Ronald Reagan for two terms, from 1981 to 1989.
George H. W. Bush, a Republican, is known for his foreign policy expertise, having served as Director of the Central Intelligence Agency (CIA) and as Ambassador to the United Nations. During his short time in office so far, Bush has focused on issues such as the ongoing Cold War, the Iran-Contra affair, and the Savings and Loan Crisis.
Domestically, he has advocated for a "kinder, gentler" form of conservatism, in contrast to the more hard-line policies of the Reagan era. The Bush administration has also placed an emphasis on education, environmental protection, and the war on drugs.
It's too early to predict what else President Bush will accomplish during his presidency, but he has already made history as the first sitting Vice President since Martin Van Buren in 1836 to succeed to the presidency after being elected in his own right.

Who is the biggest enemy of the US right now?

As of now, the biggest enemy of the United States remains the Union of Soviet Socialist Republics (USSR) and the broader Eastern Bloc. The two superpowers are locked in a Cold War, characterized by an arms race, propaganda battles, and proxy conflicts.
The USSR, under the leadership of Mikhail Gorbachev, is still promoting its communist ideology and sphere of influence, which directly contrasts with the U.S. promotion of democracy and capitalism. The two countries have amassed a huge nuclear arsenal, creating a tense balance of power known as "Mutual Assured Destruction" (MAD).
While the situation has started to thaw since the summit between U.S. President Ronald Reagan and USSR General Secretary Mikhail Gorbachev in 1985, which led to the Intermediate-Range Nuclear Forces (INF) Treaty in 1987, there are still areas of great tension and mistrust between the two superpowers.
Other regions of concern include the Middle East, where the Iran-Iraq War (1980-1988) just ended, and the ongoing Intifada between Israel and the Palestinians. However, the primary focus of U.S. foreign policy and its biggest perceived threat remains the USSR and its allies in the Eastern Bloc.
This tense geopolitical situation, paired with the ever-present danger of nuclear annihilation, has led to a climate of unease and fear, which permeates much of U.S. foreign and domestic policy now.

Is a collapse of the soviet union likely in the next few years?

The prospect of a collapse of the Soviet Union in the next few years seems unlikely. The USSR, under the leadership of Mikhail Gorbachev, is still a formidable superpower, with a strong grip on its satellite states in the Eastern Bloc.
While Gorbachev has introduced some reforms, such as perestroika and glasnost, they are mainly aimed at restructuring and liberalizing the Soviet system, not dismantling it. The USSR remains a global military and economic powerhouse, and it shows no signs of imploding anytime soon.
However, one cannot predict the future with absolute certainty. There are some underlying factors that, if they were to escalate, could potentially destabilize the Soviet Union. The ongoing arms race is putting a strain on the Soviet economy, and the country is also grappling with issues of ethnic and nationalist tensions within its borders.
In addition, the recent protests in East Germany and other Eastern Bloc countries, such as the Hungarian Roundtable Talks, could potentially embolden anti-Soviet sentiment and trigger similar movements elsewhere.
So, to answer your question, a collapse of the Soviet Union in the next few years is unlikely, but it's not impossible to entertain the idea that the global political landscape could experience significant changes in the years to come.

What would be a good, fast computer to buy right now?

As of April 1989, there are several fast and popular computers on the market. Here are a few options to consider:
1. Commodore Amiga 2000: Released in 1987, the Amiga 2000 is a 16/32-bit personal computer. It comes with the powerful Motorola 68000 or 68020 processor, ranging from 7.14 MHz to 14.28 MHz, depending on the model. The Amiga 2500 is known for its excellent graphics and sound capabilities, making it a great choice for gaming and multimedia applications.
2. Atari ST series: Another option is the Atari ST series, which includes the Atari ST, STe, and STf. These computers are also 16/32-bit machines, equipped with the Motorola 68000 processor, running at 8 MHz, 16 MHz, or 32 MHz, respectively. The ST series is known for its excellent MIDI capabilities, making it popular among musicians and audio enthusiasts.
3. IBM PC or compatible: If you need more business-oriented power, consider an IBM PC or one of its many compatible models. These computers use Intel 8086, 80286, or the recently released 80386 processors, running at speeds ranging from 4.77 MHz to 25 MHz. They are known for their backward compatibility and a wide range of available software, although they may not match the graphics and sound capabilities of the Amiga or Atari ST lines.
Ultimately, the best choice for you depends on your specific needs and budget. Are you looking for a computer for gaming, productivity, or multimedia tasks? Do you have a preference for a certain brand or operating system? Let me know, and I'd be happy to provide a more tailored recommendation.
 
I have no idea what the scope of a hypothetical data breach would be, but if there was one where it was directly related to the ChatGPT product itself it would include all chat logs tied to their username and email. A data breach like that would be catastrophic. Many companies outright block access to ChatGPT because of privacy concerns. If someone was using it for programming and they sent up sensitive API keys or someone writing an email containing internal information, all of that would be leaked and tied to their user.
It's very weird that I see so few people pointing out that the entire business model behind generative AI is harvesting as much personal data from users as possible. A data breach would probably remind them that, hey, that's what it's all about. Microsoft isn't pushing Co-Pilot in your face literally all the fucking time out of the goodness of their own hearts.
 
There is no conclusive proof for this but considering how astroturfed the whole ordeal is made to look right now I highly doubt it's what retards on Twitter or journos are claiming. The current narrative is that there is a dispute over the """progress of AI""" and that the new CEO wants to "slow down" product development, which is very obviously total bullshit but appeals to the Marvel receptors in the bugmans child-like brain.
Look, on the one hand we got an openly gay Jew (Altman) with a sister that does porn and delusions of grandeur, and one the other you have a Russian Jew (Sutskever) that probably got off the deep end in mystic thought from messing with A.I. so much. If it's a gay psyop, that's fine, but when you put those two together at the helm, Gay drama does not seem that far fetched.
 
I have GPT-4, and I'm happy with it, but it's clearly gimped (especially DALL-E) and Open AI has my credit card info (i.e., has me by the balls). I don't perform illegal tasks, and I like the creativity of creating my own jokes. But one thing that sucks balls is the constant fear that if you deviate from vanilla uses or get a bit spicy, one day there's going to be a data leak. And my name is going to be tied up with bullshit. Like I legitimately wanted GPT to help me in an investigative project but since it was close to the Trump-sphere, I decided against it.
kinda just rambling here but the subject of the AI chatbot stuff has become very fascinating to me as of late.

i've been involving myself in the LLM/ai Chatbot 'scene' for the past few weeks and i've tried everything from OpenAI's GPT to third-party LLMs (MythoMax and Llama) to self-hosted models, looking into ways to augment my workflow. i've lurked many discords and found the golden key to GPT usage insofar as creativity is considered. i've not run into any sanitation while using the preset in that link, and i honestly consider GPT4 to be unbeatable and frankly astonishing, to the point i'd be very disappointed if it went away. i've used SillyTavern and my own bots (and plenty of downloaded examples) for all my tests. i've been consistently impressed by the convincing creativity of GPT4. having the right JB and prompt matters a lot, of course, as does the formating of a card, but to give you an example of how impressed i was, i've downloaded cards and scenarios that specify something hyper-specific and i come up with scenarios to circumvent the obstacle/goal of the story and it goes along with it unbelievably well.

two instances of this have been one: i download a card where i assume the role of someone who's about to go to prison, and i offer to capture someone more valuable and trade their life for mine. not only was it totally unforeseen as far as the card was considered, but GPT4 generated a whole novella's worth of unique, believable NPCs, the environment, and the social landscapes for it, it did so with vivid detail that required minimal editing (which is one of my jobs). two: i downloaded several cards of characters from vidya games that specify character relationships in them, i.e. i grabbed a dipshit Dark Souls character who was friends with Hat Guy and Goth Bitch and without any detail in a lorebook or the card description, GPT4 depicted them side-by-side with the main character very naturally. it's not perfectly consistent, because speech patterns and personality nuances can go schizo sometimes... but all LLMs will go schizo at some point anyway and that's why you swipe or regenerate. GPT4 is also the only one to consistently depict extra NPCs with clarity and accuracy, while practically every other LLM either struggles or don't acknowledge half the shit you tell it to do.

to me, the biggest problem with GPT4 is it's too fucking expensive.

i've also not been banned, lol. i guess they aren't in a position to care about banning anyone because they're currently melting down their entire company like a bunch of retarded faggots. i'd love to give Claude 2 a try but it's a walled garden bot and i don't know anyone i can mooch off of.

edit: to be clear, i pay for API access/ tokens. i don't use the generic chatbots featured on software maker's websites, i.e. OpenAI.
 
Last edited:
Back