Best book to understand the AI phenomenon in depth, mathematical precision, and comprehensive detail?
>>24945853Kek
I'm an AI grad student. From what angle do you want to tackle this? I have read hundreds of books, papers and resources on the topic.Please tell me your background and established knowledge so that I can appropriately recommend you stuff. Also "AI" is extremely broad so you need to be more precise as to where your interest lies.Do you want to know machine learning? The computer science foundations of AI? The mathematical foundations of the algorithms we use like backpropagation, activation functions like ReLU or the workings of the modern transformer architecture?Do you instead want to know about deep learning systems and how they are trained with reinforcement learning to do things like beat everyone at chess and Go. Or do you want to know how we use adversarial networks to generate photorealistic pictures and videos?Maybe you specifically mean LLMs and how we use pre-training, fine tuning and alignment to make modern chatbots?Perhaps you mean AGI or "real artificial intelligence" and the plausible path towards it and how it could be reached?Also tell me about your goals, what do you want to achieve by reading the books or resources. Do you want a greater understanding, insight, practical knowledge or do you intend to get into the field yourself (not recommended by the way as someone working in it) I won't reply immediately but I will look back on this thread every couple of hours or so.
>>24945866Seriously though. This is where it will lead. They'll whip up some fake public consent to deny higher tech to the lower classes.But in the meantime take all this water and electrical power, when they know damn well how fusion and other ZPE devices could alleviate that waste. Just to ruin us
>>24945877This was a prompt from AI of course. No need to read it. Ask Grok what you'll think of it.
>>24945877>not recommended by the way as someone working in it)Why?
>>24945898Extremely oversaturated and overspecialized field. Meaning you will have to specialize into a subfield with a total of 20 jobs in existence that use your expertise.Its the worst discrepancy between general public expectations (AI is the future and where you make money) versus reality (AI has thousands of extremely specific specializations and none of them are hiring, one of the highest amount of grads to job opening ratios in existence)Your literature degree gives you more job security, so only go into the field out of passion. Kind of like being a writer or indie game developer.The ONLY reason I have a job at all is because I was interning at a company right when ChatGPT released and so there was a lot of investor money sloshing around. They decided to hire every intern they had working for them to pad out employee numbers to attract investors.Do yourself a favor and keep AI as a hobby and get a real job as a dentist/construction worker and have your own practice or outsourcing gig if you want to make money.I worked as a librarian for 3 years while interning into various AI company (librarian is weirdly close enough to data science so it's the closest we can get to an actual AI job)
>>24945877Masters in political science, slotted for intel work to the side of political theory hobbies. What do you know and can recommend on the future of AI integration into governance and national security?
>>24945877basically i want to build an uncensored intelligent chatbot for loneliness purposes
This will tell you how all those cool ai models work: https://keras.io/examples/In order to actually understand it you should also read some introduction to fitting neural networks, which you will find somewhere on the internet on a university's homepage. A little bit of math beyond high school level helps but isn't strictly necessary.
>>24947098If a big tech investor went out of his way to understand how these AI models work, would he short or double down on his investments?
>>24947109Even if a billionaire ai investor reads the latest ai papers for breakfast he still can't reliably predict the future of ai. Maybe it will soon replace all human intellectual work or maybe the technology will get stuck and remain the fun but not very useful curiosity it is now. I personally think that ai will be cheap and most ai startups will fail because the technology is so open and everyone can setup their own cheap service.
>>24945781Probably be faster and better to watch YouTube videos on this.
>>24947137>Maybe it will soon replace all human intellectual work Not LLMs, surely. My question was about them specifically. Are the big companies doing anything to advance AI outside of LLMs?
