2023 can be called the first year of AI, in which breakthroughs in key technologies, such as the optimization of cutting-edge algorithms, the improvement of learning Xi efficiency, and the innovation of large language models (LLMs), jointly promote the capabilities of AI to new heights. The application of AI has also achieved cross-industry penetration and diffusion, especially in key areas such as healthcare, education, autonomous driving, and supply chain management, AI technology has not only improved efficiency, but also given birth to new business models. In the face of the rapid development of AI technology, governments and regulators around the world have begun to formulate more comprehensive and mature laws and regulations to ensure the rational use of technology and risk control.
AI if it's the next Noah's Ark丨Midjourney V6
The capital market's investment in AI has reached an unprecedented level of enthusiasm, which has accelerated the pace of AI research and development and expanded its market influence. As AI technology becomes more popular, society is becoming more aware of it, and the public has a more complex and mature understanding of the life changes and potential problems brought about by AI. Combined, 2023 marks a milestone in the history of AI and the first year that AI begins to "rule the planet."
2023· Q1丨Midjourney V6
January
January 15: ChatGPT has risen rapidly to become the fastest-growing consumer app in history.
On November 30, 2022, OpenAI launched ChatGPT, an innovative AI interactive Q&A platform that allows users to ask questions and get instant feedback, as if they were having a conversation with another person. It quickly garnered a lot of attention from the public, attracting more than a million users within five days of its release. According to a report released by UBS Group AG in January 2023, the number of monthly active users of the platform has exceeded 100 million in just two months, surpassing TikTok and Instagram in record time.
January 27: Microsoft makes a major investment in OpenAI with a whopping $10 billion.
“The tech giant aims to remain at the forefront of generative artificial intelligence with its partnership with OpenAI.” （这家科技巨头希望通过与 OpenAI 的合作，保持在生成AI领域的领先地位。 ）——《纽约时报》
February
Feb. 10: Microsoft released an upgraded version of the AI-integrated Bing search engine - "New Bing".
Feb. 17: The band Linkin Park has launched the first AI-generated music video, heralding the entry of animated AI videos into the public eye.
This video uses infinite zoom technology and blends seamlessly with anime-style graphics. The AI's interpretation of the song's theme unfolds in a dynamically evolving landscape, traveling continuously, reflecting the inner struggle conveyed in the lyrics. Rely on Kaiber AI neural network to achieve.
Feb. 22: Spotify introduces an AI feature called "DJ" that aims to improve the user experience.
February 24: Amazon Web Services (AWS) and AI research firm Hugging Face announce a partnership.
February 28: Snapchat joins the ranks of AI chatbots with a new feature called "My AI".
March
March 1: OpenAI launches API services for ChatGPT and Whisper.
March 17: OpenAI released GPT-4, a more advanced language model.
AI Moments Vol.1: Understanding Pictures! Say Goodbye to Translated Texts! GPT-4 Released
To sum up, the upgraded GPT-4 is an evolution of the GPT-3 series, with more parameters and more powerful learning Xi capabilities, making it a more significant improvement in understanding and generating languages. GPT-4 can process text input, as well as image input, which in turn outputs relevant text. GPT-4 has shown performance comparable to experts in many areas of expertise and academic benchmarks, such as bar exams and providing medical advice. This means that GPT-4 has already crossed the gap between AI and human professionalism to some extent.
March 21: AI proposes new cancer treatments in just 30 days, demonstrating the huge potential of AI in the medical field.
March 22: Google releases ChatGPT rival Bard, but its stock price has since fallen.
【Ai Moment】Google Ai Dialogue Bard Open!
Like Baidu's Wenyan Yixin, although there is some room for improvement in Google's Bard, we still hope that the AI market will attract more players and achieve increasingly prosperous development. A dynamic and competitive AI market will help drive diverse competition and will inspire various players to continuously refine and optimize their technologies, resulting in smarter, more efficient, and more convenient AI tools for users.
March 22: Adobe enters the field of AI images with the release of an AI image generator called Adobe Firefly.
【Ai Moment】Getting Started with Adobe Firefly, Is the Biggest Rival of Graphic Designers Appearing?
The generation of creative words is the most surprising to me, which is also the height that other Ai painting programs have not yet reached, which can not only combine the required Prompt elements well with the text, but also achieve a reasonable construction effect and add a sense of design to creative words.
March 24: OpenAI released the ChatGPT plug-in platform, widely considered the "app store for AI".
March 24: Canva, a design software company, unveiled a new AI tool at its Canva Create campaign.
March 28: Zoom launches Zoom IQ, an AI assistant, to further enhance the online meeting experience.
March 30: Elon Musk joins AI experts in calling for a six-month moratorium on the development of more powerful AI systems beyond GPT-4 to assess societal impacts and risks.
[Ai moment] really scared or delaying time? Many bigwigs jointly called for Ai development to "slow down"!
Whether the joint letter is true or not, it is enough to sound a wake-up call for us to be in the "AI carnival" at present, and over-reliance on Ai is undoubtedly hanging the fate of mankind under the sword of Damocles, and whether it is really safe and rational to let our civilization and history take full power from Ai and make it our only scientific and technological leader?
2023· Q2丨Midjourney V6
April
April 3: Italy announces a ban on ChatGPT, sparking global concern about AI regulation.
April 6: Meta unveiled the "Segment Anything Model," an AI technology capable of extracting any object from an image or video.
April 11: Alibaba announced the launch of ChatGPT's competing product, Tongyi Qianwen.
April 14: Amazon and its cloud services arm, AWS, released Bedrock, an infrastructure service, and CodeWhisperer, an AI programming assistant.
April 17: Elon Musk launches the xAI project, although the details remain unclear.
April 21: Google's two AI research teams, Google Brain and Google DeepMind, announce a partnership.
Along the way, we've been fortunate to have two world-class research teams leading the industry forward with foundational breakthroughs that usher in a new era of AI. Progress is now faster than ever. To ensure that general-purpose AI develops boldly and responsibly, we are creating a division that will help us build stronger systems safer and more responsibly.
April 25: Artist Grimes agreed to use her voice to make AI-generated music.
April 25: AI discovers a brand new planet.
May
May 2: AI pioneer Jeffrey Hinton retires from Google and warns of the potential dangers of AI.
Hinton stressed that as AI-generated fake photos, videos and texts become more proliferating online, it is becoming more difficult for people to distinguish between false information. He also mentioned that chatbots could be more intelligent than humans and could be exploited by bad actors, for example, by automatically generating text in large quantities, making highly effective spam bots. He is also concerned that AI may replace paralegals, personal assistants, and other repetitive jobs, and believes that more jobs may be replaced in the future.
May 3: Inflection AI has released its new AI tool, Pi.
May 10: Meta has launched a new AI research model, ImageBing.
May 11: Google announces a series of major AI advancements at its annual I/O conference.
May 12: Anthropic expands the window for its AI model Claude to 100,000 users.
May 19: OpenAI released its free ChatGPT iOS app.
OpenAI has released its ChatGPT mobile app
Similar to its desktop version, the ChatGPT app provides users with the opportunity to interact with an AI chatbot to ask questions, get advice, find inspiration, or use it for Xi and research without having to go through a traditional web search.
May 19: Apple bans its employees from using ChatGPT.
Why Apple Rarely Mentions "Ai"
Maybe one day, when the Ai hype subsides, Apple will redesign Siri to make her look and voice more like ChatGPT. At that time, all of Apple's machine Xi and neural network technology will converge into a striking "one more thing", and Apple will also promote it in a tone that avoids emphasizing Ai as much as possible. However, for Apple at this stage, keeping a low profile and focusing on user privacy and security are still its core competitiveness.
May 24: Microsoft unveils Windows 11 AI Copilot, Bing plugin, Bing browser, integrated AI.
June
June 2: NVIDIA Research has developed Neuralangelo, a tool capable of 3D reconstruction from 2D video clips.
June 13: An AI religious service was held in a church in Germany with more than 300 participants.
June 19: Meta released Audiobox, a new text-to-speech AI technology.
June 22: More than 100,000 ChatGPT account information was compromised, raising security and privacy concerns.
June 27: Groundbreaking AI research projects have announced that they have been able to understand and decode the communication between whales.
JUNE 28: BAIDU LAUNCHED ITS AI MODEL, ERNIE 3.5.
June 29: ChatGPT launches its mobile search service through Microsoft's Bing search engine.
In the second quarter of 2023, AI technology has achieved rapid development in many fields such as regulation, R&D innovation, product launch, and application, and has also sparked extensive discussions about its safety and ethics from all walks of life. The progress made during this period shows that AI technology is permeating all aspects of human life, and it will play an increasingly important role in people's daily work and life.
2023· Q3丨Midjourney V6
July
July 3: Humane unveils the "AI Pin," a portable device designed to replace smartphones.
The AI Pin is a wearable device that interacts through voice, gestures, and a built-in projector, connects to an AI model (such as GPT-4) to provide functions such as communication, information query, and translation, and automatically processes user requests through the Cosmos operating system, without the need for traditional application management.
July 4: Midjourney has improved its travel and adventure AI services with the launch of a new "Panning" feature.
July 6: OpenAI introduces the concept of super alignment to solve the problem of super smart alignment.
July 10: Research shows AI shows potential in treating brain cancer.
July 12: Anthropic unveils Claude 2, an upgraded version of its AI model Claude.
GPT-4 supports more languages and is suitable for global applications, while Claude 2 generates text more naturally. GPT-4 is strong in code generation, and Claude 2 is good at code optimization. GPT-4 has a wide range of tasks and integrates search, while Claude 2 is suitable for in-depth document analysis. GPT-4 has strong comprehension and reasoning skills, and Claude 2 has better writing, math and coding skills.
July 17: Hollywood strikes over the use of AI in film and television production.
July 19: Meta and Microsoft jointly release Llama 2, a new AI model.
Qualcomm & Meta have cooperated on mobile LLMs
July 26: Japan's Ministry of Economy, Trade and Industry announced plans to develop an AI supercomputer.
August
Aug. 7: Apple CEO Tim Cook says the company is incorporating AI technology into all of its products.
Apple CEO Tim Cook said that AI and machine learning Xi technology are deeply integrated into almost all of Apple's products and are the key to its success. Although AI was not mentioned at the recent developer conference, Cook confirmed that Apple has been developing these cutting-edge technologies for many years and continues to increase its investment in the field. This shows that Apple is not lagging behind its competitors and that AI is still at the heart of its product design. Cook emphasized that Apple has long put AI and machine Xi at the core of its strategy, and will continue to increase R&D investment to further improve the AI capabilities of its products and services.
August 11: HeyGen announces version 2.0 of its AI-generated avatar cloning technology.
August 23: Meta open-sourced an AI model called SeamlessM4T.
Aug. 24: Researchers use brain implants and AI to enable paralyzed patients to communicate.
Aug. 30: Tesla unveils a new $300 million AI supercomputer.
Tesla plans to deploy a powerful supercomputing cluster with 10,000 Nvidia H100 GPUs, which can reach 340 FP64 PFLOPS, one of the strongest in the world. The cluster will significantly enhance Tesla's AI training capabilities and accelerate the development of its fully self-driving technology. At the same time, Tesla is also developing its own custom supercomputer, codenamed Dojo, with an estimated investment of more than $1 billion. Dojo's enhanced performance with specially optimized chips will synergize with Nvidia clusters, giving Tesla computing power unprecedented in the automotive industry
September
Sept. 4: Researchers have created an AI model capable of predicting the smell of chemicals.
September 8: Microsoft has partnered with Paige to develop a new cancer detection AI.
Paige leverages its vast digital pathology database, which contains more than 500,000 slice samples totaling about 1 billion images. These images cover multiple types of cancer, providing rich training data for deep Xi models. Big data sets can allow AI models to learn more about the characteristics of cancer Xi, improving the accuracy and generalization of diagnosis. The Paige team's extensive expertise in digital pathology ensures that AI models don't just focus on the generic features of images, but also identify and interpret subtle differences that are critical to cancer diagnosis and treatment.
Sept. 13: Coca-Cola uses AI to create a new mystery flavor.
September 20: Google launches the Bard Workspace extension, designed to boost productivity.
September 21: OpenAI released DALL-E 3, which further improves image generation technology.
September 22: Microsoft unveiled "Your Everyday AI Companion" called Copilot.
September 26: ChatGPT enables multi-modal recognition of speech and images.
【AI Moment】GPT has long eyes!Epic features are quietly released!— OpenAI releases GPT-4V multimodal model
After experiencing regulatory restrictions, computing power tensions, AI ethics discussions, and the pursuit and interception of new competitors, in the early morning of September 26, Beijing time, OpenAI quietly released its GPT-4V model, that is, a multimodal model, the most eye-catching of which is the video recognition function
September 27: The CIA builds a proprietary AI system for Intel.
September 28: OpenAI announces the return of ChatGPT's browsing capabilities.
September 29: Mistral AI launched an open-source 7B level language model.
In the third quarter of 2023, AI technology continued to be widely used in multiple fields, achieving major breakthroughs in business growth, healthcare, entertainment, enterprise services, and more. Driven by competition and cooperation in the business world, AI technology has developed rapidly, and at the same time, it has also aroused the attention of all sectors of society to its impact, such as Hollywood's resistance to new technologies. As AI becomes more integrated into various industries, its ethical, employment, and legal implications will be discussed more broadly.
2023· Q4丨Midjourney V6
October
October 3: Rewind unveils the Rewind Pendant, a wearable device that captures real-world conversations.
October 4: LinkedIn showcased its new AI tool across products.
Oct. 13: Meta's AI character chatbot goes live, but is met with backlash.
October 16: AI deciphers the first text in the burned Herculaneum Scroll.
November
November 6: Elon Musk's xAI launches Grok, an AI chatbot based on X.
11月7日：OpenAI在DevDay上展示了GPT Builder、GPT-4 Turbo、Assistants API等。
"I'm going to go Open up!
GPTs allow anyone to create a customized version based on ChatGPT, which can be applied to different scenarios such as daily life and professional work, and build their own AI assistant without programming. Users can input specific knowledge into GPTs, set the scope of use, and then share them with others.
Nov. 9: Amazon invests in training a large language model codenamed "Olympus".
Nov 14: OpenAI confirms it is developing GPT-5 and advanced AI agents.
11月16日：Forward Health推出AI医疗舱。
Nov. 17: YouTube and Google DeepMind unveiled a new AI model called "Lyria."
Nov. 17: Sam Altman is fired as CEO of OpenAI.
November 22: Sam Altman is rehired as CEO of OpenAI.
November 29: Pika launches Pike 1.0, a text-to-video generator.
Nov. 30: Google DeepMind's "GNoME" generates formulations for new inorganic compounds.
December
December 1: Microsoft invests $3.2B in UK AI infrastructure.
12月6日:Google DeepMind发布Gemini。
The Gemini series includes three models: the lightweight Gemini Nano, the powerful Gemini Pro, and the high-end Gemini Ultra. Gemini Nano is designed to run natively offline on Android devices, Gemini Pro will power many of Google's AI services and is the underlying model for the new product Bard, and Gemini Ultra is the most powerful language model Google has ever created for data center and enterprise applications.
Out of 32 benchmarks with OpenAI's GPT-4, Gemini outperformed 30, especially when it came to processing video, audio, and multimodal data. Gemini is also more computationally efficient than Google's earlier models, based on the self-developed TPU v5p chip. Google plans to apply Gemini to its major products such as search, advertising, and Chrome browser, which is seen as the key to the future development of Google's AI.
December 12: French startup Mistral AI launches Mixtral.
December 13: Microsoft Research releases Phi-2.
December 20: Microsoft partners with Suno AI for AI song generation.
December 22: Apple research showcases AI technology for iPhones.
December 24: Midjourney V6 beta update
[AI Moment] This time it's too real, it's really a bit infringing!—Midjourney V6 update experience
V6 brings significant improvements to the data source and algorithms to produce more realistic, detailed, and detailed images. It has achieved in-depth optimization in the portrayal of materials, portraits, and light and shadow effects, improving the overall texture and realism of the image.
However, more and more people are discovering that the reason why Midjourney V6 is able to produce extremely realistic drawings may be because its training source has penetrated into every frame of the film and television work, and it does not hide the consistency of the raw pictures. Midjourney V6 creates animated and comic-book-style images that closely match the original work in great detail, exposing the blurred boundaries of current AI painting technology when it comes to copyright compliance. At present, there is no specialized agency responsible for reviewing the copyright legality of AI training datasets, which provides companies with room to operate in the regulatory blank area of copyright law.
December 27: OpenAI is in talks to raise additional capital at a 100B+ valuation.
December 28: The New York Times sues Microsoft and OpenAI for copyright infringement.
At the end of 2023, AI technology is booming in many aspects, and its influence is penetrating into social media, military, healthcare, content creation, and other fields. At the same time, discussions around AI ethics, security, and intellectual property are becoming increasingly intense. The growing investment and R&D activities of companies and research institutes in this area underscore the positive expectations for the future development of AI technology and its potentially significant impact on the global economy.