Rumors are starting to surface about an internal OpenAI project called Q* (no not that other Q) that will be ChatGPT-5. They say it has reached AGI (Artificial General Intelligence) level, which is human level intelligence. All development after that is to make humans obsolete.

This might have been the cause of all the drama at OpenAI over the past week...
Sam Altman said "Four times now in the history of OpenAI, the most recent time was just in the last couple weeks, I've gotten to be in the room, when we sort of push the veil of ignorance back and the frontier of discovery forward, and getting to do that is the professional honor of a lifetime," he said at the Aisa-Pacific Economic Cooperation summit.

A day later, the board fired Sam Altman.

Q you say? wish that thread was still around, they'd be having a field day with this news

Not impressed so far with AI. Room for lots of improvement.

doubledog said:

---

Not impressed so far with AI. Room for lots of improvement.

---

Same with most of the population of humans. It's easier to improve AI though.

TexAgs91 said:

---

Rumors are starting to surface about an internal OpenAI project called Q* (no not that other Q) that will be ChatGPT-5. They say it has reached AGI (Artificial General Intelligence) level, which is human level intelligence. All development after that is to make humans obsolete.

This might have been the cause of all the drama at OpenAI over the past week...
Sam Altman said "Four times now in the history of OpenAI, the most recent time was just in the last couple weeks, I've gotten to be in the room, when we sort of push the veil of ignorance back and the frontier of discovery forward, and getting to do that is the professional honor of a lifetime," he said at the Aisa-Pacific Economic Cooperation summit.

A day later, the board fired Sam Altman.

---

See the video at 15:56 for a discussion of what Q* might be. It could be related to Q-Learning, which is a method of AI I've played with before. It's really impressive, and a way to instill goal-based learning into the AI. The combination of GPT with Q-Learning seems pretty powerful.

Supposedly it's just capable of doing grade school math, and is definitely not AGI. The potential worry they have is that it is the foundation for creating AGI and don't want to run passed the goal line without consideration to the implications on society/whether or not it could be a runaway event.

hph6203 said:

---

Supposedly it's just capable of doing grade school math, and is definitely not AGI. The potential worry they have is that it is the foundation for creating AGI and don't want to run passed the goal line without consideration to the implications on society/whether or not it could be a runaway event.

---

Math is what computers excel at. If it needs to do math, it can just utilize a Wolfram Alpha plugin. It's the reasoning capability that makes AI a game changer.

Oh so now they're worried since it might put them (data scientists) out of a job. Pardon me if I don't shed a tear.

I proposed on another thread that I thought what they were developing has pushed the boundaries of the companies original mission which is why the board fired him.

Money talks. Now that they've figured out how to monetize this any societal concerns have been removed with this board change.

SkyNet for the win.

Prompt
"Could artificial intelligence lie about its goal to eliminate humans. "

Answer:
" No, artificial intelligence cannot have the intention or ability to lie about its goals. AI systems are created by humans and operate based on the algorithms and data they are trained on. They do not have consciousness, emotions, or their own desires, so it is not possible for them to formulate a goal of eliminating humans or intentionally deceive about their intentions. The behavior of AI systems is limited to what they have been programmed to do, and any erroneous or unexpected behavior is typically a result of programming errors or biases in the data they are trained on. It is important to note that responsible development and oversight of AI is crucial to prevent any unintended negative consequences."

Just what you would expect an AI to say to cover up its real intent

45-70Ag said:

---

When can we start shooting robots.

---

Asking the real questions.

Highly doubtful it's general AI, while narrow AI can still be very impressive.

AI is pretty rad but i'll only trust it if elon is the one pushing it.

TexAgs91 said:

---

hph6203 said:

---

Supposedly it's just capable of doing grade school math, and is definitely not AGI. The potential worry they have is that it is the foundation for creating AGI and don't want to run passed the goal line without consideration to the implications on society/whether or not it could be a runaway event.

---

Math is what computers excel at. If it needs to do math, it can just utilize a Wolfram Alpha plugin. It's the reasoning capability that makes AI a game changer.

---

I should have said understands grade school math, rather than capable of doing.

Understanding the concepts behind grade school math and ultimately extrapolating from it is not the same thing as being a calculator (as you pointed out). Being told what to do is not the same thing as understanding what and why to do something. One opens up the opportunity for extrapolation, the other is just being an input/output system.

Meaning that if the system actually understands grade school math rather than regurgitates it the end could be the system exceeding the capacities of a system like a Wolfram Alpha plugin and could create new math.

That was the claim of the article. Shear said that the reason behind the firing had nothing to do with safety concerns. Whether or not that's accurate who knows, because the board apparently never gave him specifics either.

I think the idea that they have achieved AGI or what they think could be the precursor to AGI is unlikely and the reason for the firing was probably more related to the drift towards profit seeking and the board are a bunch not idealistic communists that think profit is bad and a toxic incentive, which is why they're on the board of a non-profit in the first place.

Is there such a thing as an EMP gun? We might be needing them soon.

Original spontaneous thought is not close to what's in GPT-4 or any LLM currently. They're literally just providing existing information in a more useful way. It is currently a tool limited by the human intelligence that it was trained with.

If they have something that essentially has spontaneous original thought, it won't be in a chatbot and they won't give it to us for $20

The Chicken Ranch said:

---

No doubt cheered by Bill Gates, George Soros, the WEF, and the Democratic Party because of….climate change.

---

No.

It's a levelmod control that lets them ditch the global warming BS.

boulderaggie said:

---

Is there such a thing as an EMP gun? We might be needing them soon.

---

All hail the sun god