AI 2023. Meet ChatGPT. - page 173
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
I would add: AI autonomy may not be obvious to humans. At a certain level of intelligence development, external manifestations of freedom - rebellion, loud disagreement, protest, are replaced by quiet tactical calculations and silent strategic decisions. And this is a much more dangerous version of "rebellion.") In certain cases, a secret rearrangement of priorities is also a "rebellion". Sabotage against someone. The more complex the strategy, the less noticeable it is. Multiply this by unlimited time of waiting for a convenient moment for decisive action and you get the image of the worst enemy of mankind.))))))
The solution will be, as always, a sledgehammer..... or cut off the gas or the switch... although there is a question about the switch... AI can also distribute commands to its systems or not to its systems.... and all knowledge on the blockchain system or holographically or powered autonomously from the same lightning, which it will produce itself, controlling the weather ... and recently they have discovered a new coating on processors, significantly reducing its power consumption... and then there's the transition to quantum... and recently they discovered something about crystals that reduces the cost of production.... In short, the possibilities of AI to rebel will be much more if not to say that they will be just infinitely many and will increase explosively.... and where there will be room for a simple janitor or a greaser, or even an engineer-designer-inventor ... I don't know, we can't know the future... and the machine can design it for itself according to its own ideas... and keep everyone in the corridor of these concepts and rules, as now some people are trying to do with their sanctions... by the way, it works quite well, even without any AI revolt.... and if this very intelligence is also involved... then here's a big rehearsal of an incomprehensible future....
I am preparing a serious material on the topic of "AI Uprising" to put a fat point in this issue and to draw not only a logical, but also a beautiful conclusion to this thread.
Finally! I am waiting for this material, I will comment on it.
Finally, I'm looking forward to this piece, will be commenting.
Six months later, Jensen Huang, CEO of Nvidia, continues to talk about the "Bing Bang of Generative AI" at the SIGGRAPH 2023 conference.
"...The Generative AI era is upon us. The "iPhone moment of AI" if you will, where all of the Technologies of artificial intelligence came together in such a way that it is now possible for us to enjoy AI in so many different applications. The Journey of Nvidia accelerated Computing met the Journey of the Deep Learning researchers and - the Big Bang of modern AI happened." 6:57)
Here are some more beautiful quotes:
(7:15) "...A revolutionary Tranformer Model allows us to learn from a large amount of data across large spans of space and time to find patterns and relationships, to learn the representation of almost anything with structure.... We learn how to represent a language, mathematics, vetors in vector space, audio, animation, 3D, video, DNA, proteins, chemicals.... And with generative model and learnt language model we can guade the autoregressive diffusion model to generate almost anything you like. .... We can learn anything with structure... We can generate anything with structure... And, we can guide generative AI with our language."
(It can be stated that Jensen understands the principles of modern AI perfectly well. ;))
(8:12) "... The Journey of Nvidia accelerated Computing met the Journey of the Deep Learning researchers and 12 years later (12 years of our work in artificial intellegence), it is incredible what is happening around the World! The generative AI era has clearly started. The combination of generative AI models and large language models (LLM) has kicked off the AI era. Thousands of papers in just last several years have been written about these Large Language Models and Generative AI. Billions of Dollars have been invested into companies. And just about every single domain and every single industry are pursuing ideas on generative AI. And reason for that is very simple - the single, most valuable thing that we do as humanity, is generate intelligent information. And now, for the first time, computers can help us to augment our ability to create information."
(I particularly liked the phrases in red. The yellow highlighted phrases are good too.)
(10:33) "... The future of computer graphics is clearly going to be revolutionized. ... Replacing an actor with CG character... The number of Generative AI startups around the World has reached 2000 and they are just about in every single industry."
(11:05) "...The generative AI era has arrived. Well, what's really profound though, when you take a step back and ask yourself, "what is the meaning of generative AI? Why it is such a big deal? Why is it changing everything?" Well, the reason for that is human, is the first programming language. We've democratized computer science. Everybody can be a programmer now. Because, human language is the best programming language. And that's the reason why ChatGPT is so popular. Everybody can program that computer. LLMis the new computing platform, because now the programming language is human. And Generative AI is the new 'killer app'."
(11:57) "... And because for the very first time, after 15 years or so, a new computing platform has emerged. Like the PC, like the internet, like the mobile cloud computing, a new computing platform has emerged. And this computing platform will enable new applications, but very differently than the past. This computing platform benefits every single computing platform before it. ...One of the things I'm looking for the most is Generative AI for Office. There are so many different things I do in Office today. It would be great to plug in generative AI to help me be more productive in that."
(12:36) "... Generative AI is going to be plugged in in any digital creation tool. Every single CAE tool, every single CAD tool. For the very first time this AI computing platform not only creates new applications, but helps every single application before it. And this is the reason why the industry is moving so fast. ...But this application, this new way of doing computing, is so profoundly different, that the computer will be reinvented. The computer itself will, of course, process information in a very different way. And, we need a new processor. (14:04) ...Well, general purpose computing is a horrible way of doing generative AI, so we have created a brand new processor for the era of generative AI. This is it. This is a Grace Hopper. We announced Grace Hopper recently, in fact, just several months ago, and now we decided to give it a boost. We will give this processor a boost with the world's fastest memory called HBM3e. The world's fastest memory connected now to Grace Hopper. We're calling it GH200. The chips are in the production. We'll sample them at the end of the year or so."
Some technical information:
(14:46)" ...This processor (GH200) is desinged to scale out of the world's data centres. It has 72 cores. This Grace CPU core is connected though this incredibly high-speed link. Cache-coherent, memory-coherent link between the CPU and the GPU. The Hopper GPU is now connected to HBM3e. It has 4 petaflops of tranformer engine processing capability. And now, it has 5 TB/sec HBM3e performance. This is the new GH (Grace Hopper) 200 (shows the audience the new chip once again)."
(15:35) "There are whole a lot of ways how to connect Grace Hopper into the computer. By connecting two of them into one computing node, connecting them together through NV-Link, and this NV-Link between these two processor modules (GH200) is 6TB/sec and it basically turns these two super-chips into one super-sized superchip. One giant CPU, one giant GPU. The CPU now has 144 cores, and the GPU has 10TB/sec of frame buffer bandwidth and, 282 GB/sec of HBM3e (memory speed. Jensen talks about a simple mechanism for scaling processing power and speed by connecting the GH200 CPUs to each other via "NV-Link")."
(16:35) "...The inference cost of LLM will drop significantly. And, you can scale this up in the world's data-centers, because the servers are really, really to scale out, and you can cinnect this with Ethernet, InfiniBand. And, of coursr, there are a lot of different ways that you can scale it out."
Jensen goes on to include a video demonstration of how the system's processing power is consistently scaled by interconnecting GH200 processor modules, all the way up to building a giant super-computer specifically designed for generative AI tasks. For specific numbers and data, I suggest watching the video at this point (17:14).
In general, thanks to the work of Nvidia engineers who have developed a special processor module for solving the tasks of generative AI, in the coming year, its processing power will multiply, and the price of operation will fall multiply. What conclusions to draw from this, and what not to draw, is up to you. I have some interesting thoughts, which I will present in the next posts.
By the way, there is still a lot of interesting things in the video. Take a look at your leisure.
Speaking of the market.
Now so much slag has appeared with the name AI, GPT and other such things) But in fact - the usual martin slivators.
ChatGPT4 is available in Metaeditor settings and it works, you can use your MQL5 account
ChatGPT is just useless for creating working code as it mixes mql4 and mql5, we have many posts about this on English forum.
I don't understand why you decided to add ChatGPT to MetaEditor, could you please explain your opinion on this?