Speed is constant in the world of Artificial Intelligence: many important things have happened this week too, I'll tell you about three technological ones (Chat GPT Code Interpreter, Google Bard and Claude AI) plus a sadder one!
1) OPEN AI Code Interpreter: the beauty of speaking in code
OPEN AI has released the new 'Code Interpreter' mode in beta for ChatGPT Plus users. A tool that is the basis of the programming revolution for almost everyone.
Quoting Open AI in its first alpha release presentation :
[Code Interpreter is]... an experimental ChatGPT model that can use Python, handle uploads and downloads. We provide our models with a working Python interpreter in a sandboxed, firewalled execution environment, along with some ephemeral disk space. Code run by our interpreter plugin is evaluated in a persistent session that is alive for the duration of a chat conversation (with an upper-bound timeout) and subsequent calls can build on top of each other. We support uploading files to the current conversation workspace and downloading the results of your work.
Why is it important?
In a nutshell, it provides an easy-to-use way to combine the use of Large Language Models (LLM) such as GPT, with more traditional programming skills, without the need to write code or set up an environment to run it. The thing that makes it interesting is that it allows incredible analysis even for non-experts.
Currently the documentation is poor and the links appear limited (You cannot access the graphs or download the files). But if you want an idea, while suffering from the lack of graphs, click here.
Data scientist what do you think, is your job in danger too?
2) Google Bard: a new arrival in Italy.
Google Bard has finally been available in Italy for a few days without the need to use a VPN.
I confess that while it has much broader nominal capabilities than GPT, Bard still leaves me very puzzled by its contrived language, the results I'm getting, and what might be Google's goals in that regard.
The characteristics that distinguish it are:
1) it is possible to make him describe images by providing the link
2) make requests by voice
3) share conversations and export the generated code.
I'll study it calmly, in the meantime you can start using it too and get an idea.
Why is it important?
Google is finally coming out after long months since the release of GPT Chat with an important product. OpenAI therefore has a first important competitor (the next one is Antropic, we'll talk about it in the next news). Will Big G be able to create a serious solution this time?
3) Antropic releases Claude 2.0 - The Newcomer
Another great awaited novelty in the world of Large Language Models.
Antropic is a San Francisco-based company founded in 2021 by the Amodei twins, who previously served OpenAI and left the company due to "differences of opinion" after Microsoft joined. The company, with only 7 employees, has raised 1 .5 billion dollars of investments in a short time to develop an AI Chatbot based also on Transformers.
To test it outside USA or UK you must use a VPN (for example by installing the Opera browser which has one integrated) and point to https:/ /claude.ai
Why is it important?
A new player with very clear ideas, many means available that with version 2.0, a bit paternalistic in the answers and still not connected to the Internet, is able to accept incoming files and has a conversational window of 100,000 tokens (basically it fits well more than an entire book, well over 5 hours of reading).It seems to give very consistent answers and is poised to be a great alternative to ChatGPT. We'll see!
p.s but what is a Context Window?
At the end of the article I'll explain it along with a few other code words.
+1) Start the impact of AI in the real world?
Sad but true: Suumit Shah, the CEO of an e-commerce platform, announced that to lay off 90% of their support team due to chatbot innovations AI since responses to customers are provided in two minutes instead of two hours.
Why is it important?
Yesterday I was at a wonderful workshop in which one of the main concepts that emerged was that AI is not a technology problem.
In other words, the technology will run its course and solve problems as they arise, but we must prepare for the potentially profound impacts that AI could have on the real world.
Quick Glossary
GAI (Generative Artificial Intelligence): A type of artificial intelligence that can generate new and original content.
OPENAI: An artificial intelligence research organization that aims to promote and develop human-friendly AI in a safe and beneficial way.
Python: A high-level, readable and easy to learn programming language.
LLM (Large Language Models): AI-based language models that can generate human-like text at a large scale.
Token: In this context, a token can be a word or part of a sentence that a language model uses to understand and generate text.
Context Window: The amount of tokens a language model can "remember" to generate consistent responses.
Transformers: A deep learning technology used for natural language understanding.
Chatbot AI: A software program that can conduct a conversation via text or audio.
Comments