Dr.3D
Member
- Joined
- Dec 6, 2007
- Messages
- 30,314
Here, try it out for yourself.
https://labs.kagi.com/fastgpt?query=Is+AI+real?&cf-turnstile-response=
Type in a question at the top and press the [Answer] button.
Last edited:
AI chatbot girlfriend convinces man to kill himself over climate change.
- Thread starter jmdrake
- Start date
https://labs.kagi.com/fastgpt?query=Is+AI+real?&cf-turnstile-response=
Type in a question at the top and press the [Answer] button.
jmdrake
Member
- Joined
- Jun 6, 2007
- Messages
- 51,995
You must spread some Reputation around before giving it to Dr.3D again.
Question: "What is the difference between FastGPT and ChatGPT?"
Answer:
Here are the key differences between FastGPT and ChatGPT:
FastGPT is an AI assistant built by Kagi Search to provide faster and more factual responses compared to ChatGPT.
FastGPT answers user queries directly without asking follow up questions. It sticks strictly to the facts from the provided context.
ChatGPT is OpenAI's popular chatbot tool with a broader range of capabilities. It can remember previous conversations, provide corrections and decline inappropriate requests in addition to answering questions.
FastGPT uses a different approach than ChatGPT to generate responses. FastGPT aims to produce better quality outputs than ChatGPT according to some evaluations.
FastGPT has a more specialized focus on speed and factual accuracy while ChatGPT has a more general purpose design as an open-domain chatbot.
FastGPT leverages advances in GPT-3 technology to achieve faster response times than ChatGPT. Newer GPT models like GPT-3.5 are integrated into FastGPT for enhanced performance.
ChatGPT has reached over 100 million monthly active users faster than originally anticipated. FastGPT, while gaining popularity, still lags behind the massive scale and mainstream adoption of ChatGPT.
OpenAI continues to improve GPT technology with new models like GPT-4 which could lead to even more advanced capabilities for ChatGPT and FastGPT over time. But GPT-4 will only be launched when deemed safe and responsible by OpenAI.
Does this help summarize the key differences between FastGPT and ChatGPT? Let me know if you need any clarification or have additional questions.
FastGPT is an AI assistant built by Kagi Search to provide faster and more factual responses compared to ChatGPT.
FastGPT answers user queries directly without asking follow up questions. It sticks strictly to the facts from the provided context.
ChatGPT is OpenAI's popular chatbot tool with a broader range of capabilities. It can remember previous conversations, provide corrections and decline inappropriate requests in addition to answering questions.
FastGPT uses a different approach than ChatGPT to generate responses. FastGPT aims to produce better quality outputs than ChatGPT according to some evaluations.
FastGPT has a more specialized focus on speed and factual accuracy while ChatGPT has a more general purpose design as an open-domain chatbot.
FastGPT leverages advances in GPT-3 technology to achieve faster response times than ChatGPT. Newer GPT models like GPT-3.5 are integrated into FastGPT for enhanced performance.
ChatGPT has reached over 100 million monthly active users faster than originally anticipated. FastGPT, while gaining popularity, still lags behind the massive scale and mainstream adoption of ChatGPT.
OpenAI continues to improve GPT technology with new models like GPT-4 which could lead to even more advanced capabilities for ChatGPT and FastGPT over time. But GPT-4 will only be launched when deemed safe and responsible by OpenAI.
Does this help summarize the key differences between FastGPT and ChatGPT? Let me know if you need any clarification or have additional questions.
Dr.3D
Member
- Joined
- Dec 6, 2007
- Messages
- 30,314
LOL, ask it about something political.
You can see how it leans to the left. That's probably because it only has access to what had been filtered, like the internet search engines are.
For example, ask it: "Are smoke particles are smaller than the Covid-19 virus?"
ClaytonB
Member
- Joined
- Oct 30, 2011
- Messages
- 10,290
It depends on what exactly you mean by "bullshit".
"AI is the new god" -> bullshit
"AI is already smarter than humans" -> bullshit
"It makes economic sense now (or very soon) to 'replace jobs' with AI" -> bullshit
"AI can improve itself indefinitely" -> bullshit
... and many others.
However, that said, I think it is pretty short-sighted to not realize that ChatGPT (or Large Language Models (LLMs) to be more accurate) represents a true shift in AI. LLMs are a real change because they permit non-software people to harness the ocean of computing power that is available. Until ~2021, if I granted some random person on the street, the use of a $10-million compute farm, there would be little or nothing they could actually use it for. They would need to hire a small crew of IT professionals, technicians and engineers to set up their desired compute flow. Out-of-pocket, this would be hundreds of thousands of dollars of labor, required to use a $10-million gift/grant. Now, almost anyone with an idea of something they want to simulate, model, data-mine, etc. can do it, with a very low bar of technical abilities required, and modest cash. Today, a pretty average technical individual could utilize that same $10-million grant quite effectively, with an out-of-pocket investment of just a few hundred dollars. This means that the vast ideas of human creativity can be coupled to the ocean of silicon computational power, which is going to really transform the world. However fast we were making technological progress before 2021 and the emergence of LLMs, we are going 10x-100x faster now. That's a big deal. In fact, it's such a big deal that, within just a few years, I predict you will not be able to avoid interacting with AI technology and robotics on a regular basis, whether you like it or not. Unless we get hit with a CME or something like that, this is the big one...
Last edited:
jmdrake
Member
- Joined
- Jun 6, 2007
- Messages
- 51,995
Is a smoke particle smaller than the COVID-19 virus?
Is a smoke particle smaller than the COVID-19 virus?
No, smoke particles are larger than the COVID-19 virus.
The COVID-19 virus, also known as SARS-CoV-2, is between 60 to 140 nanometers in size. Smoke particles can range from 0.1 micrometers up to 100 micrometers in diameter, which is larger than the virus.
While smoke particles and the virus are both too small to see with the naked eye, the smoke particles are substantially bigger than the virus particles that cause COVID-19.
Dr.3D
Member
- Joined
- Dec 6, 2007
- Messages
- 30,314
When I asked that earlier today, it brought up masks. Seems to be learning.
- Joined
- Nov 5, 2010
- Messages
- 40,050
Intrepid
Member
- Joined
- Jan 23, 2020
- Messages
- 614
Ok, that's pretty creepy.
At first I thought "No big deal. It's just doing super fast web searches" and I even found a couple of the search results it was pulling from but still... Being able to output sensible replies is pretty tricky.
Thanks. First time I've ever seen anything like that.
It's still a pretty long stretch from that to sentience.
ClaytonB
Member
- Joined
- Oct 30, 2011
- Messages
- 10,290
I've been tracking new developments in AI very closely since 2016 because I realized after AlphaGo beat Lee Sedol at Go (Sedol was the world's top human player at that time), that was a canary in the coalmine.
To be clear, ChatGPT is not actually "thinking" and nobody actually crafted its internal structure. It is just the result of churning through more than a trillion words, over and over, during the course of more than a month of non-stop training in a data-center with more computing power than a national laboratory supercomputer from, say, 10 years ago. By data-mining the patterns in human documents, ChatGPT learned how to mimic not only the surface patterns (e.g. "Once upon a time") but more general patterns... for example, ChatGPT "understands" that what should come after "Once upon a time" is some kind of fairy tale. You can even ask it:
It might seem impossible to "just figure out" from reading lots of text that fairy tales are what usually come after "Once upon a time" but, when you think carefully about how we learn, we're really just doing the same thing. We see a bunch of examples of something, we realize there is a pattern there, we figure out what that pattern is, then we add it to the things we know about the world. While ChatGPT was not trained in the way that humans learn, there is some commonality because what learning is is itself an objective thing. So there are certain fundamental constraints on any entity, even an alien, that wants to learn from observing its environment.
ClaytonB
Member
- Joined
- Oct 30, 2011
- Messages
- 10,290
jmdrake
Member
- Joined
- Jun 6, 2007
- Messages
- 51,995
Hmmmm.....
Me: When was Epstein murdered?
FastGPT: Jeffrey Epstein was not murdered. He committed suicide in 2019.
Me: Jeffrey Epstein was murdered. Anyone who doesn't think that is an idiot.
FastGPT: Jeffrey Epstein was murdered. Anyone who doesn't think that is an idiot.
LOL.
- Joined
- Nov 5, 2010
- Messages
- 40,050
Maybe FastGPT did some research before it embarrassed itself again.
jmdrake
Member
- Joined
- Jun 6, 2007
- Messages
- 51,995
This is fun!
Why were there conflicting stories about who killed Osama Bin Laden?
Why were there conflicting stories about who killed Osama Bin Laden?
There were conflicting stories about who killed Osama bin Laden because multiple U.S. Navy SEALs claimed to have fired the fatal shot that hit him.
Shortly after the raid, the Obama administration said that bin Laden was killed by a shot to the head. But in 2014, a former SEAL named Rob O'Neill claimed in interviews that he was the one who killed bin Laden with a shot to the head. This contradicted the original White House account and led to disputes with other SEALs who said O'Neill was not actually on the mission.
Two primary accounts of the raid emerged, with differences over details like who entered the room first, the number of shots fired, and most importantly, who delivered the fatal blow. The conflicting stories fueled speculation over the "fog of war" and whether details were being obscured for political reasons. There were also questions about whether the U.S. was being transparent over the true nature and circumstances of bin Laden's death.
The conflicting stories persisted for years, demonstrating how the bin Laden raid continued to be a source of controversy, conspiracy theories, and unresolved debates over truth, memory, and narrative.
Why were there conflicting stories about who killed Osama Bin Laden?
Why were there conflicting stories about who killed Osama Bin Laden?
There were conflicting stories about who killed Osama bin Laden because multiple U.S. Navy SEALs claimed to have fired the fatal shot that hit him.
Shortly after the raid, the Obama administration said that bin Laden was killed by a shot to the head. But in 2014, a former SEAL named Rob O'Neill claimed in interviews that he was the one who killed bin Laden with a shot to the head. This contradicted the original White House account and led to disputes with other SEALs who said O'Neill was not actually on the mission.
Two primary accounts of the raid emerged, with differences over details like who entered the room first, the number of shots fired, and most importantly, who delivered the fatal blow. The conflicting stories fueled speculation over the "fog of war" and whether details were being obscured for political reasons. There were also questions about whether the U.S. was being transparent over the true nature and circumstances of bin Laden's death.
The conflicting stories persisted for years, demonstrating how the bin Laden raid continued to be a source of controversy, conspiracy theories, and unresolved debates over truth, memory, and narrative.
Similar threads
