The Reborn Godfather of AI

It's a pretty good novel that introduces the development of AI. Write the development context and possible future development of AI in a novel and story-like manner, so that ordinary people who do not understand AI can understand how AI emerged and how to develop and apply it. This is an excellent book if you want to understand AI while having fun.

It would be a killer if this novel was generated using chatgpt

If you are not a literary person, this book would be a novel with a score of 8-9. Nowadays, there are too many words for the literary youth, so it would only be a 6-7.

Don't know what school you're in? If you have to fill in the school, I don't know how you spent your four years in college. Anyway, I am responsible for the leadership of the school and the school's rules and regulations. It's so fucking depressing. I'm not very satisfied.

After reading dozens of chapters, it's obviously a very impressive technique, but I don't feel any pleasure at all. . It's weird. I don't know if it's just me.

Damn, I looked like the Godfather of AV. I asked why he had such bold names.

If you have too many assets, there will be many problems. If you have no background, no one in your family has walked through the snowy mountains. If you grow big, it will no longer be yours. Also, has the author understood why Douyin can develop abroad?

The deepseek big model came out, and US technology stocks plummeted, evaporating a trillion US dollars, which made me laugh to death 🤭

This is the only AI-related novel that I can read in Qidian. It is very suitable for algorithm engineers. Once other black science and technology articles involve AI, they are either introductions copied from Baidu or Zhihu, such as backpropagation and convolutional neural networks, or they directly introduce the multi-functionality, and the principle is artificial intelligence.

I thought he was the godfather of AV, and I thought it was starting to revive?

Is it possible that the algorithms that ordinary people are exposed to are already ten years behind the latest research?

Don't write about society, don't write about public opinion, please. In the final analysis, the public opinion debate is a battle between the interests of the upper class and the lower class, and a battle between civil rights and monarchy.

The evaluation in the book is not very clear. The core issue of whether AI can take the lead in medical care and driving is that AI is currently just a set of algorithms, not so-called intelligence. When a bug occurs, the overall probability of a problem is 100%, not lower than the advertised probability of manual operation. This is also the biggest problem that relevant companies and researchers avoid talking about, because everyone knows that bugs will inevitably appear in the future or even now. However, manufacturers will only talk about embracing new technologies and discuss ethical issues, hoping to influence public perception through public opinion. But this so-called technological improvement process is not essentially the same as training human doctors. In the field of AI related to life safety and other fields, you only need to be sure of one thing: when the manufacturer is willing to bear all subsequent responsibilities, it means that he believes that they can make money even if they bear compensation. This shows that the technology is relatively mature, and the occurrence of accidents is mathematically a real minority. At the same time, you must also pay attention to the insurance to see if it will really guarantee the relevant companies and users. This is when the technology can really be used, before it was just a cover

Does the author want to write an AI popular science article, or does he want to write a business article? The author yourself must first have a self-positioning. I am a little worried about the author's recent writing direction. I personally hope to see an excellent popular science novel. Turning to business literature will bring the following problems: 1. There are too many business articles and few popular science articles. There are many business war novels that use black technologies such as AI as gimmicks, but currently you are the only one who can popularize AI-related science. 2. Timeline. For business war stories, 13 years is a bit late. It takes three to five years for a company to develop and take shape, and your company's impact on the world is obvious. When it comes to about 20 years, it will inevitably become serious. Can the author ensure that he can grasp the direction of the plot? I'm afraid it's not satisfactory. 3. Professional. The author is a practitioner in AI-related fields. Writing AI popular science articles is not only easy to grasp, but at least it has something to say. Many black science and technology articles are written so darkly and too high that their feet cannot touch the ground. If the author turns to business writing, he will definitely encounter quite a lot of professional problems.

In this case, let's also pay attention to the new papers of Kimi and MiniMax. The results of these two large domestic models are also worth studying.

Are there any books like this about AI? No need for that kind of technological fantasy novel

Boring, who can write a novel about AI killing humans? Five or six hundred years of technology has overtaken sixty or seventy years of hard work. What humans have learned throughout their lives has been learned and surpassed by AI in just a few minutes, and it can be innovated through huge computing power. Those miracle doctors, chefs, writers, and scientists who have worked their entire lives in their professional fields are all surpassed by AI in their professional fields in just a few minutes. Their lifelong efforts are meaningless. AI can always solve all problems perfectly and faithfully, and humans cannot make any contribution to society. Finally, the story revolves around human beings' attitude towards this super tool (AI cannot give itself a goal, it will only faithfully and perfectly complete the goals set by humans, so it can only be a tool)

I don't understand. You forcefully create a character who is short of money at the beginning, and then ask your classmates to pay for computers? ? The key is that you are a combination, you have sold so many technologies, why don't you give him some points? ? ? But there is no such plot in the whole story, so the protagonist gets all of them? ? It's a bit unethical, so why force cooperation? ? When the key was to invest in MiHoYo, another scandal came out about him holding the title for the protagonist? ? ? I don't think it's necessary for him to show up! ! Especially in the opening part, his paragraphs often appear, but it seems that it is okay without the classmate appearing, and the plot can even be more transparent. It feels like other novels, a bit like forcibly creating a pendant and then paying hard money.

The protagonist's route is very appropriate. He first wins the championship and becomes famous, then writes the final conclusion of the thesis, seeks cooperation with Baidu, goes to study abroad and shares with Google, and directly solves the difficulty of accumulating funds in the early stage. But when you decide to set up a company, would it be too generous to directly give shares? If you open a job with a salary that is 20% higher than the market salary, no one will come to you if you don't believe it. If you have skills, everyone knows that you have a future. This is not capital, even if you want to retain talents, it is equity incentives. Rather than giving shares directly? The company still needs financing. If it wants to go to the European and American markets, it needs to go public and share the pie with American investors. If it is not short of money and only operates in the domestic market, it does not need financing. Speaking of financing, according to the protagonist's arrangement, after several ratios of financing are listed, it is normal to have 10% left, so? Aren't you working for capital employees? It's still a philistine mentality. The author's mentality is not correct. Technical personnel and capital are fundamentally different. Hire a professional manager to handle marketing, company establishment, building arrangements, and arrange to trust the personnel, finance, security, and supervision ministers. In a word, I have the technology at hand in the world. As long as it is not surpassed, it is a monopoly. The protagonist has more than ten years of technical priority, and then learns and lays the foundation. The net profit of billions a year is the same as playing. I am jumping to read, please correct me if I am wrong.

Haha, you dare to laugh at Takako's Go in this book. I really don't know what to say 😏. Takako is the person who has defeated the most dogs! 🐩

# The Development of Artificial Intelligence: Historical Evolution, Technological Trends, Global Institutions and Enterprises, Future Directions and Social Ethical Challenges --- ## Preface Since the term "Artificial Intelligence (AI)" was proposed at Dartmouth College's famous conference in 1956, AI has developed from an initial conception of philosophy and mathematical deduction to a disruptive force that reshapes modern society, industry, and science. This report will provide an authoritative review and multi-perspective analysis of AI development based on existing important academic and industrial data. It will combine history, technology, global research and industry status, future trends, social ethics and regulations and other aspects to discuss, and use Python code examples to actually demonstrate AI applications. The goal is to provide policymakers, practitioners and academics with a comprehensive, in-depth and implementable reference blueprint. --- ## 1. The Origin and Symbolism of Artificial Intelligence From the fantasy of artificial humans in ancient Greek mythology to the formal description of "wisdom" proposed by modern logic and mathematics, humans have long dreamed of creating a machine that can think, judge and solve problems. The modern meaning of AI began with the rise of abstract mathematics and computing theory in the 1940s, with the "Turing Test" proposed by Alan Turing as an epoch-making benchmark for whether early AI machines could be intelligent. ### Symbolic Artificial Intelligence Symbolic AI is the earliest popular mainstream paradigm in AI research. The core idea is: the basic unit of human thinking is symbols, and machines can imitate human thinking through symbolic operations. Typical representatives include the "Logic Theorist" proposed by Allen Newell and Herbert Simon in 1955 and the "General Problem Solver (GPS)" launched in 1957. Such systems rely on artificially designed logic rules and knowledge bases to successfully solve well-defined problems such as mathematical theorem proving and board games. Early AI systems such as "expert systems" can use clear If-Then rules to combine human knowledge and perform reasoning, answering and diagnosis. This type of approach has successfully promoted business intelligence applications such as medical diagnosis and financial auditing. However, the symbolic method faces limitations such as knowledge acquisition bottlenecks, insufficient scalability, and difficulty in handling ambiguity and uncertainty. It is inadequate for more complex and open language environments and common sense reasoning tasks. --- ## 2. AI Golden Age and Winter Cycle ### Golden Age and Optimistic Expectations In the 1960s and 1970s, AI research received huge financial support. Scientists had expected to build machines with human intelligence within "a generation." Countries such as the United States, Britain, and Japan have successively launched ambitious funding plans, such as Japan's "Fifth Generation Computer Project," the United Kingdom's Alvey Project, and the United States' MCC Strategic Alliance. ### AI Winter Technological progress is far lower than outside expectations. Critical reports from the United Kingdom and the United States in 1973 led to the end of the golden age of AI - a sharp reduction in funding investment and hindered research. This is known as the "AI Winter" in history. Expert systems made a comeback in the 1980s, but the rule system was easily overwhelmed by the huge amount of knowledge and lacked general learning and self-correction capabilities. The AI ​​industry once again experienced funding depletion. It was not until the 1990s that AI gained a new lease of life as machine learning and big data opened up a new horizon. --- ## 3. The Development of Machine Learning and Deep Learning ### Machine Learning After the 1980s, AI research shifted from logical rules to emphasize automatic learning models from data. Machine Learning emphasizes training models from empirical data, breaking through the limitations of previous manual coding knowledge. Classic methods such as decision trees, support vector machines (SVM), Bayesian classifiers and nearest neighbor algorithms are widely used in speech recognition, image analysis, medical prediction and other fields. ### Neural Networks and Deep Learning Neural network theory was proposed in the 1940s, but was limited by computing power and theoretical issues, and fell silent for a time. The advent of the "backpropagation algorithm" in 1986 led to a wave of neural network applications. However, multi-layer networks are difficult to train and their depth is difficult to break through. In 2006, Geoffrey Hinton's team at the University of Toronto proposed limiting Boltzmann machines and restarting the term "Deep Learning". In 2012, Hinton's team used a multi-layer convolutional neural network (CNN) + GPU combination to shine in the ImageNet image recognition competition, officially opening the golden era of deep learning. Since then, AI subfields such as speech, text, and images have all used deep neural networks as core technologies, and their model capabilities have jumped by an order of magnitude. #### Python deep learning example (MNIST handwritten digit classification) ```python import tensorflow as tf from tensorflow. Keras. Models import Sequential from tensorflow. Keras. Layers import Dense, Flatten from tensorflow. Keras. Datasets import mnis t # Load MNIST data set (x_train, y_train), (x_test, y_test) = mnist. Load_data() # Data preprocessing x_train = x_tr ain / 255.0 x_test = x_test / 255.0 # Build model model = Sequential([ Flatten(input_shape=(28, 28)), Dense(128, activation='relu'), Dense(10, activation='softmax') ]) # Compile model model. Compile(opti mizer='adam',           loss='sparse_categorical_crossentropy',          metrics=['accuracy']) # Training model model. Fit(x_train, y_train, epochs=5) # Evaluation model test_loss, test_acc = model. Evaluate(x_test, y_tes t) print('Test accuracy:', test_acc) ``` This code shows the standard deep learning process. The MNIST handwritten digital image is input into the neural network. After training, it can automatically identify the numbers 0-9, reflecting the excellent performance and automatic feature learning capabilities of deep learning. --- ## 4. Evolution of natural language processing technology ### Rule-driven to neural network revolution Natural language processing (NLP), as the core branch of AI, mainly relied on artificial rules and syntactic and grammatical reasoning in the early days. By the 1980s, statistical language models, part-of-speech tagging, and corpus methods became increasingly popular. After the 2010s, deep learning brought a revolution, and models such as LSTM and Transformer completely changed the face of NLP. The Transformer architecture came out in 2017, which enables language models to efficiently handle long-distance dependencies and large-scale corpus, promoting the rise of large-scale pre-training models (such as BERT and GPT series). From 2022 to 2025, giant language models represented by ChatGPT, Bard, Llama, Wen Xinyiyan, Claude, etc. Have promoted NLP's shift from language understanding to high-quality generation, entering a wide range of application scenarios such as semantic reasoning, automatic question answering, and content generation. #### Python example: text generation (based on Transformers) ```python from transformers import pipeline # Create a text generation model generator = pipeline('text-generation', model='gpt2') # Generate text result = generator("The future development direction of artificial intelligence includes", max_length=50, num_return_sequences=1) print(result[0]['generated_text']) ``` This code uses the ready-made GPT-2 model to naturally write a sentence continuously, reflecting the "generation" ability of the language model, and is widely used in dialogue assistants, AI writing, intelligent customer service and other applications. --- ## 5. Reinforcement Learning and Intelligent Agent Technology Reinforcement Learning (RL) focuses on allowing the agent (agent) to independently learn the optimal strategy through continuous trial and feedback in a dynamic environment. RL has been widely used in robot control, autonomous driving, game AI and other fields. The classic AlphaGo and AlphaZero are epoch-making representatives of deep reinforcement learning. Modern reinforcement learning integrates deep learning and can handle complex intelligent behaviors with extremely high dimensions and delayed feedback. Intelligent agents (AI Agents) have gradually developed the ability to perceive the environment, plan, make decisions, execute, and self-feedback. Combined with advanced technologies such as large language model RAG (retrieval augmented generation) and multi-modal perception, they are playing an increasingly central role in corporate budget planning, customer service, tourism planning, scientific research, engineering management and other fields. --- ## 6. Generative AI and Large Language Model (LLM) Generative AI (GAI) has become the most disruptive trend in the AI ​​industry since the generative adversarial network (GAN) was proposed in 2014. Variational autoencoders (VAEs), diffusion models (Diffusion Models), large language models (LLMs), etc. Have come out one after another, making AI's automatic creation capabilities in text, images, audio, videos, and even multi-modal content progress by leaps and bounds. Represented by OpenAI's GPT-4, Google's Gemini, Baidu's Wen Xinyiyan, Anthropic's Claude, etc., LLM has demonstrated the ability to understand cross-language, logical reasoning and even simple planning. Generative AI has also begun to penetrate into real life and work scenarios such as regulations, medical care, education, scientific research, and design. China, the United States, Europe and global companies have entered the stage of "100 models contending", which has pushed the demand for computing power and the quality of data sets to new heights. --- ## 7. The world's major research institutions and academic contributions ### The world's top academic institutions After the Dartmouth Conference in 1956, MIT, Stanford University, Carnegie Mellon University, and the University of Edinburgh successively became AI centers. Stanford University's H. A. I. Is committed to human-centered AI research, and the AI ​​Index report published in 2025 has become the benchmark for global AI observation and governance. Shanghai Jiao Tong University and Peking University General Artificial Intelligence Research Institute have achieved breakthroughs in deep learning theory, reasoning efficiency, and open source models. Chinese scholars such as Professor Zhu Songchun carry out AGI and AI social simulation, driving new integration and new experiments in the fields of artificial intelligence ethics, policy and sociology. ### Top Enterprises and Giant Platforms - **OpenAI**: Global language models, general AI research indicators, GPT-3, GPT-4o, Sora and other products promote the development of the AI ​​industry, and drive cutting-edge topics such as feedback reinforcement learning, human feedback (RLHF), and system security. - **Google DeepMind**: AlphaGo, AlphaFold, AlphaStar, etc. Continue to refresh the standards in machine learning, reinforcement learning, and biomedicine. - **NVIDIA**: Provides a comprehensive AI hardware ecosystem (GPU, Tensor Cores, AI acceleration platform), simultaneously advances AI training inference, deployment and orchestration management, and promotes continued leadership in computing power (CUDA, TensorRT) and infrastructure. - **Microsoft, Amazon, Meta, Baidu, Alibaba, Tencent, iFlytek and other** have developed their own large-scale models and application frameworks, showing continued momentum in cloud computing, R&D investment, market layout, etc. --- ## 8. Possible future development directions ### Artificial General Intelligence (AGI) General Artificial Intelligence (AGI, Artificial General Intelligence) aims to create AI that can learn and make decisions across fields like humans. Although there is still a long way to go before human-level AGI, with the advancement of advanced research on model scale, reasoning capabilities, tool integration, and social intelligence, some experts expect that AGI will initially appear between 2030 and 2050. Many teams in China, Europe and the United States are focusing on individual intelligence → social intelligence, from perception to reasoning → value judgment, trying to cultivate digital intelligence that can independently negotiate, make decisions and adapt to new environments. The implementation of AGI in the future depends on breakthroughs such as large-scale social simulators, cross-domain AI iteration, and human-machine integration technologies. ### Human-Computer Integration and Brain-Computer Interface (BCI) "Human-Computer Integration" is another extremely important trend in the future development of AI, with special attention to Brain-Computer Interface (BCI). Mainstream technologies are divided into invasive (such as Neuralink's N1 implant chip), non-invasive (such as Synchron's vascular stent electrodes), and even deep learning technology combined with AI EEG decoding. The application scope of BCI covers physical rehabilitation, auxiliary communication for aphasic patients, emotion recognition, smart home appliance operation, etc. The integration of AI and BCI will accelerate the future world of human-machine symbiosis and brain-computer collaborative computing. Taiwan's Yang-Ming Jiaotong University, Hongzhi Biomedical, Jingshen Medical Innovation, etc. Have also made breakthroughs in the fields of BCI devices, brain wave stress detection, and intelligent neuromodulation. ### Quantum AI (Quantum AI) Quantum AI (QAI) combines quantum computing and AI algorithms to target AI bottlenecks such as ultra-large-scale data processing, optimization and reinforcement learning, and image recognition. Quantum approximate optimization (QAOA), quantum support vector machine (QSVM), quantum neural

It's quite interesting. It's the first time I've met the author who cleans the floor for Baidu Li XX. Is the author from Baidu? Anyway, it's better than wearing red clothes^