SlackGPT - Slackforce is using "the force"
I think this is a significant step in the right direction with regards to practical application in the enterprise space. What remains to be sufficiently addresses are data security implications. However, I suspect that as more and more open source models become viable and reliable, the option of including an offline or inhouse LLM in place of a public API become available.
A lot of the focus and hype surrounding AI is around generative models and the ability to create and generate new data. There are a number of examples of this from the recent OpenAI GPT-2 model which was found to be able to generate news articles which were indistinguishable from those written by humans.
More recently, the OpenAI CLIP model has been released which is able to classify images based on their text descriptions. The model has been trained on 400 million images with their corresponding labels, in order to learn a mapping between text and images. The model is able to classify any image into a label and also generate a text description of the image.
However, while such models are exciting and novel, they are not the most practical or useful application of machine learning and AI. These models require large amounts of data to train on and are very costly in terms of compute resources, which is prohibitive for small businesses.
领英推荐
Instead, the potential benefit of machine learning is the ability to make predictions using data that you already have. For example, if you are a small business owner, you may have a large amount of sales data – but perhaps not enough to be able to generate new sales data.
However, you could use machine learning to predict future sales based on your existing data. This is much more useful for businesses than being able to generate new data. For example, you could use machine learning to predict how many customers you will have in the next month, or to predict how much revenue you will have next month.
As a result, I think the most useful application of machine learning and AI will be to predict new data based on existing data.
https://venturebeat.com/ai/salesforce-launches-slackgpt-embracing-generative-ai-for-enterprise-workflows/