In August 2020, I wrote about the stunning storytelling prowess of another LLM, GPT3 (bit.ly/3RbHfbB). The Generative Pre-trained Transformer Version 3, I wrote, was being heralded as the first step towards the holy grail of AGI (Artificial General Intelligence), where a machine has the capacity to understand or learn any intellectual task that a human being can. GPT has been trained on a massive body of text, mined for statistical regularities or parameters or connections between different nodes in its neural network. The scale is gargantuan, with 175 billion parameters; all of Wikipedia comprises just 0.6% of its training data! GPT-3 was developed by OpenAI too, and with DALL-E, it took this to another level.

In my last article, we saw that artificial intelligence is getting better and better at answering our questions, regardless of the subject matter or the sector. Using GPT-3 technology, I demonstrated that an AI can successfully pass an SEO multiple-choice test as well as solve SEO case studies. There are other, even more advanced, technologies, such as Deepmind's Gopher, which outperform GPT-3 in the following fields: Humanities, social sciences, medicine, science, and math. The following graph highlights the accuracy of the answers provided by Gopher, UnifiedQA, GPT-3, and a human expert. Depending on the subject, we can see the narrow gap between the level of the AI and that of an expert.

Since GPT-2, there has been much excitement around the applications of large language models. And in the past few years, we've seen LLMs used for many exciting tasks, such as writing articles, designing websites, creating images, and even writing code. But as I have argued before, there's a wide gap between showing a new technology do something cool and using the same technology to create a successful product with a workable business model. Microsoft, I think, just launched the first real LLM product with the public release of GitHub Copilot last week. This is an application that has a strong product/market fit, has immense added value, is hard to beat, is cost-efficient, has very strong distribution channels, and can become a source of great profit.

AIがAIに関する学術論文を書いてしまいました。書けてしまいました。AIに論文執筆をさせたのは、スウェーデンのヨーテボリ大学の博士研究員、Almira Osmanovic Thunström氏。AIによる論文完成までの行程が、Scientific Americanにて公開されています。

The service has a new responsible AI system that filters out harmful content and helps detect abuse. Additionally, Azure OpenAI Service now offers access to more models, including GPT-3, Codex and embeddings models. Codex can generate code and translate plain language to code, while embeddings make semantic search and other tasks easier. The service also offers new capabilities for customers to fine tune models for more tailored results. Azure OpenAI Service is enabling customers across industries from health care to financial services to manufacturing to quickly perform an array of tasks.

AI is making huge strides in the global industries. Business leaders from all over the world are striving to deploy AI and leverage the benefits that this technology to ensure that they gain an edge over their competitors. Currently, AI researchers and engineers are busy developing self-sufficient artificial intelligence systems. Their next step is to attain artificial general intelligence, which will enable AI systems to perform without any human supervision and also compete with human intelligence. And it seems like scientists are quite close to attaining an academically smart AI system. Swedish researcher Almira Osmanovic Thunström describes vividly how her team initially conducted the experiment by asking GPT-3 to write an academic paper about itself.

You may be wondering if such a bold headline is true. GPT-3 came out in 2020 and established a new road the whole AI industry has been following in intention and attention since. Tech companies have repeatedly built better, larger models, one after another. But although they've put millions into the task, none of them has fundamentally changed the leading paradigm or the game's rules GPT-3 laid out two years ago. Gopher, Chinchilla, and PaLM (arguably the current podium of large language models) are significantly better than GPT-3 but they are, in essence, more of the same thing.

Deep image classification models are typically trained in a supervised manner over a large, annotated dataset. Although a model's performance will improve as more annotated data becomes available, large-scale datasets for supervised learning are often difficult and expensive to obtain, requiring numerous hours of effort from expert annotators. With this in mind, one may begin to wonder if cheaper sources of supervision exist. Put simply, is it possible learn high-quality image classification models from data this is already publicly available? The proposal of Contrastive Language-Image Pre-Training (CLIP) model [1] -- recently re-popularized due to its use in the DALLE-2 model--by OpenAI answered this question in a positive fashion.

In my last article, we saw that artificial intelligence is getting better and better at answering our questions, regardless of the subject matter or the sector. Using GPT-3 technology, I demonstrated that an AI can successfully pass an SEO multiple-choice test as well as solve SEO case studies. There are other, even more advanced, technologies, such as Deepmind's Gopher, which outperform GPT-3 in the following fields: Humanities, social sciences, medicine, science, and math. The following graph highlights the accuracy of the answers provided by Gopher, UnifiedQA, GPT-3, and a human expert. Depending on the subject, we can see the narrow gap between the level of the AI and that of an expert.

Artificial intelligence (AI) can devise methods of wealth distribution that are more popular than systems designed by people, new research suggests. The findings, made by a team of researchers at UK-based AI company DeepMind, show that machine learning systems aren't just good at solving complex physics and biology problems, but may also help deliver on more open-ended social objectives, such as the goal of realizing a fair, prosperous society. Building a machine that can deliver beneficial results humans actually want – called "value alignment" in AI research – is complicated by the fact that people often disagree on the best method to resolve all kinds of things, and especially social, economic, and political issues. "One key hurdle for value alignment is that human society admits a plurality of views, making it unclear to whose preferences AI should align," researchers explain in a new paper, led by first author and DeepMind research scientist Raphael Koster. "For example, political scientists and economists are often at loggerheads over which mechanisms will make our societies function most fairly or efficiently."