Completely agree. Every new tool has the "cycle," but what is AI? It's this generation's calculator, and yes, Miss Smith (my 3rd grade math teacher, I do have a calculator in my pocket and on my wrist), so now this generation will need to figure out how to use AI as we did with calculators. (if that makes sense)
AI is Automated Intelligence, not artificial Intelligence. AI is the process of embedding human knowledge into our computer systems to reduce the complexity of functioning in an ever more complex world of expanding technology.
JJ you nailed that correctly. I am not a computer ai specialist just an Indiana to I would say. Love ai thanks to all the videos made available from you an and the developer team at IBM and RUclips Crowdcast in the past was awesome. The part you nailed was the information in the second brain. We all have them in our brain. The computer is just smarter than us at making the information work for us. I’ve been waiting for this moment for several years as this augmented intelligence has been based on the information we feed into the system. As a simple tradesperson the value I hope to infuse back to my sector or fellow trades person lies in the information I would deem factually accurate. The way about getting it in their hand is the challenge. You all have been doing a great job in developing the tools to do this and hope one day I too will be able to share my piece of my investment into something better to help myself and the people that want or need to know. Thanks
Artificial Librarian.... 👍 Great analogy for LLMs, helpful, that said there are some DL systems that do not fit cleanly this descriptor like AlphaFold.
I'm wondering if what your guest said about the AI at the beginning of the conversation is actually accurate and here is why. I understand that the guest is an AI dev. I used to write my own neural networks too so I'm aware of the tech and math behind it. Also I'm aware of different architectures. However it's not true that the model looks into the database to fetch information you ask for. As you know, the knowledge of the model is in the entire neural network (in the weights to be precise). So there is no actual database if we are not using the RAG. It's very similar (although not the same) to how mammal's brains work. I also cannot agree with te statement that "there is no reasoning". If you had a lot of different kinds of conversations with LLM then you must notice the reasoning. OK, yes, you can call it "putting different kinds of information together" but if you do that, then you will have to notice that mammal's brains do the same thing. In general: being aware of the tiniest bits of code behind LLM doesn't mean that you draw a conclusion it's just cogs working together because the same way microbiologist could say that in human brain there are only dumb cells and they can't add up to collectively make reasoning happen. Anyway, the rest of the interview including the open source and Instruct Lab is really interesting and I'm going to have look into it. Thanks! I love watching your series! 😊
Yep, you are correct. There is no "database," but when attempting to make AI make sense to everyone, I speak in metaphors to help people understand and build off of. So they can gain the most understanding in a short amount of time for us to learn together. Thank you for pointing that out, and I'm excited to see you join our InstructLab community!
@@jjasghar I understand where you are coming from. Although I think that for people new to the field of AI it's important to understand what it is from the very beginning because otherwise it will be hard for them to get the proper understanding later. Metaphors have their place, but at the same time it is worth being careful not to simplify too much, which can sometimes lead to misunderstandings. I had many conversations with people unfamiliar with neural networks and they were all telling me that AI is just an advanced database which is matching your query to it's knowledge and fetching the answer for you. I had to explain to all of them that it's not like that and there is no fetching happening. Instead of fetching there is something that you could call 'realization', and there is no database but all of the knowledge is stored in the entire network at once and there is no way to pinpoint where exactly given information is located in the network because it potentially is in the entire network at once. Anyway, I hope you don't take my input as a criticism and I don't want to come across as arrogant. I see it more as a debate if you will. Thank you for your earlier reply BTW 😊
We should not have to look for information in analog data storage systems or do mindless repetitive tasks anymore it's a waste of energy, time, and money 😎🤖
You chose to get a guy on here, to answer this question, who is so biased he won't even use the term "AI"? I love how he draws the line at saying "Intelligence" despite initially referring to it as "not that smart". If you are going to virtue signal, at least be consistent. I wonder if he calls them "mobile celluar telephone devices" instead of "smartphones" too. Such a bold and brave individual taking such a stand.
In these conversations, you need to set a framework to build a valuable discussion. Sometimes, it requires you to reframe words to make sure you bypass topics that won't gain the ultimate goal of such a complex topic. I'm sorry you had a problem with the words I used to help people grasp the most of this complex topic. And honestly? I call my phone an "iPhone." Does that fit?
I'm glad to hear something a bit more realistic for once and not AI will be our savoir/doom.
Completely agree. Every new tool has the "cycle," but what is AI? It's this generation's calculator, and yes, Miss Smith (my 3rd grade math teacher, I do have a calculator in my pocket and on my wrist), so now this generation will need to figure out how to use AI as we did with calculators. (if that makes sense)
AI is Automated Intelligence, not artificial Intelligence. AI is the process of embedding human knowledge into our computer systems to reduce the complexity of functioning in an ever more complex world of expanding technology.
JJ you nailed that correctly. I am not a computer ai specialist just an Indiana to I would say. Love ai thanks to all the videos made available from you an and the developer team at IBM and RUclips Crowdcast in the past was awesome. The part you nailed was the information in the second brain. We all have them in our brain. The computer is just smarter than us at making the information work for us. I’ve been waiting for this moment for several years as this augmented intelligence has been based on the information we feed into the system. As a simple tradesperson the value I hope to infuse back to my sector or fellow trades person lies in the information I would deem factually accurate. The way about getting it in their hand is the challenge. You all have been doing a great job in developing the tools to do this and hope one day I too will be able to share my piece of my investment into something better to help myself and the people that want or need to know. Thanks
Thank you for these kind words, and I'm looking forward to seeing you help out your sector...Open Source work isn't just for computer nerds!
Thank you for letting us know what an InstructLab is. I'm joining thanks to you JJ.
Looking forward to it! ❤
I love IBM contents! Lots of useful info - latest info, thanks guys!
Glad you enjoyed it!
Thank you for sharing the thoughts.
Artificial Librarian.... 👍
Great analogy for LLMs, helpful, that said there are some DL systems that do not fit cleanly this descriptor like AlphaFold.
Great point!
AK in Action from Zambia, Africa
I'm wondering if what your guest said about the AI at the beginning of the conversation is actually accurate and here is why.
I understand that the guest is an AI dev. I used to write my own neural networks too so I'm aware of the tech and math behind it. Also I'm aware of different architectures. However it's not true that the model looks into the database to fetch information you ask for. As you know, the knowledge of the model is in the entire neural network (in the weights to be precise). So there is no actual database if we are not using the RAG. It's very similar (although not the same) to how mammal's brains work. I also cannot agree with te statement that "there is no reasoning". If you had a lot of different kinds of conversations with LLM then you must notice the reasoning. OK, yes, you can call it "putting different kinds of information together" but if you do that, then you will have to notice that mammal's brains do the same thing.
In general: being aware of the tiniest bits of code behind LLM doesn't mean that you draw a conclusion it's just cogs working together because the same way microbiologist could say that in human brain there are only dumb cells and they can't add up to collectively make reasoning happen.
Anyway, the rest of the interview including the open source and Instruct Lab is really interesting and I'm going to have look into it.
Thanks! I love watching your series! 😊
Yep, you are correct. There is no "database," but when attempting to make AI make sense to everyone, I speak in metaphors to help people understand and build off of. So they can gain the most understanding in a short amount of time for us to learn together. Thank you for pointing that out, and I'm excited to see you join our InstructLab community!
@@jjasghar I understand where you are coming from. Although I think that for people new to the field of AI it's important to understand what it is from the very beginning because otherwise it will be hard for them to get the proper understanding later. Metaphors have their place, but at the same time it is worth being careful not to simplify too much, which can sometimes lead to misunderstandings. I had many conversations with people unfamiliar with neural networks and they were all telling me that AI is just an advanced database which is matching your query to it's knowledge and fetching the answer for you. I had to explain to all of them that it's not like that and there is no fetching happening. Instead of fetching there is something that you could call 'realization', and there is no database but all of the knowledge is stored in the entire network at once and there is no way to pinpoint where exactly given information is located in the network because it potentially is in the entire network at once.
Anyway, I hope you don't take my input as a criticism and I don't want to come across as arrogant. I see it more as a debate if you will. Thank you for your earlier reply BTW 😊
That’s how an LLM works. An LLM is one type of AI.
Hey I'm from Zambia I would like to join
We should not have to look for information in analog data storage systems or do mindless repetitive tasks anymore it's a waste of energy, time, and money 😎🤖
I think so
You chose to get a guy on here, to answer this question, who is so biased he won't even use the term "AI"?
I love how he draws the line at saying "Intelligence" despite initially referring to it as "not that smart". If you are going to virtue signal, at least be consistent. I wonder if he calls them "mobile celluar telephone devices" instead of "smartphones" too. Such a bold and brave individual taking such a stand.
In these conversations, you need to set a framework to build a valuable discussion. Sometimes, it requires you to reframe words to make sure you bypass topics that won't gain the ultimate goal of such a complex topic. I'm sorry you had a problem with the words I used to help people grasp the most of this complex topic. And honestly? I call my phone an "iPhone." Does that fit?