TOP 5 AI news in November 2022
November was a busy month in the AI world, with a number of major news stories. From investments in AI research to the release of cutting-edge AI-powered products, the month was full of exciting innovations.
In this blog post, we'll take a look at the top 5 AI stories that made waves in November and explore the implications for the future of AI technology.
GALACTICA - Meta AI language model trained on a novel high-quality scientific dataset called NatureBook, which consists of 48 million papers, textbooks and lecture notes, millions of compounds and proteins, scientific websites, and encyclopedias totaling 120 billion parameters.
Galactica is great for summarizing academic literature, solving math problems, generating Wiki articles, writing scientific code, annotating molecules and proteins, and more.
The model a lot of attention and criticism from the community, words: “dangerous”, and “biased”ecause of that, 3 days, it was introduced Meta AI pulled down Galactica without explanation.
More information: https://galactica.org/
Great video about GALACTICA: https://www.youtube.com/watch?v=ZTs_mXwMCs8&t=1s
CICERO - November Meta AI presents another outstanding AI achievement. They develop the first AI model to achieve human-level performance in the popular strategy game Diplomacy. CICERO achieved more than double the average score of human players and ranked in the top 10 percent of participants. This is a fantastic achievement Diplomacy viewed as a near-impossible task for AI for most of the decade. is because AI needs to understand other people’s motivations and perspectives and have an excellent strategic mindset.
More information: https://ai.facebook.com/blog/cicero-ai-negotiates-persuades-and-cooperates-with-people/
Stable Diffusion 2.0 & Midjourney v4 - The two best diffusion models released new versions, and the results are stunning. Stable Diffusion's new version presents new features like higher resolution images and depth-guided transformations. Midjourney team was focused on realistic imagery. Midjourney is way ahead in generating images but Stable Diffusion is open source and free, thus is used to create more interesting applications.
More information about Stable Diffusion 2.0: https://stability.ai/blog/stable-diffusion-v2-release
More information about Midjourney version 4: https://midjourney.gitbook.io/docs/
GitHub Codespaces - November GitHub organized Satellite Conference, an annual presenting recent innovations. One particular one was Codespaces, a cloud-based IDE accessible from a standard web browser. It provides all essential tools for software development without cluttering local devices. Main features: Instant dev environment, developing to debugging all at a single place, easy code contributions with coworkers. One cool feature is that you can access your Jupyter Notebook from GitHub.
More information: https://github.com/features/codespaces
OpenAI GPT-3 text-davinci-003 - OpenAI released a new version of GPT-3. Major improvements by OpenAI:
• Higher-quality writing.
• handle more complex instructions, meaning you can get even more creative.
• Longer-form content generation
However, the new capabilities of Davinci also require more computing resources leading to higher average social media post costs, around 6 cents.
More information: https://openai.com/api/
Thank you for reading, and see you next month! And if you think where is a ChatGPT, we will write about this next month!
Author: Deividas Matačiūnas