Articles about AI
Facebook wants to kill ChatGPT

Facebook wants to kill ChatGPT

Benoît's picture

Updated on Aug 25, 20233 min read

Table of contents

image of street art with Mark Zuckerberg (taken by Annie Sprat)

Some days ago, we learned that ChatGPT could faces bankrupt in 2024. And yesterday, Facebook introduced a tool that could kill ChatGPT. It is free and better than ChatGPT

Let's dive in 👇

Introducing Code Llama: The ChatGPT Killer

In a bold move, Meta Platforms has released Code Llama, a powerful artificial intelligence (AI) tool aimed at revolutionizing the world of coding. Designed to generate computer code, review, fix programming language, and offer support in several languages such as Python, C++, Java, PHP, Typescript, and Bash, Code Llama is poised to become a major competitor against OpenAI's Chat GPT.

The free downloadable large language model (LLM) comes in different sizes, catering to various needs for quicker responses or more thorough output. With this release, Meta asserts its dominance in the development of artificial intelligence tools by offering a free chatbot specifically designed to assist with writing computer code.

A Closer Look: What Can Code Llama Do?

Code Llama is built on top of Llama 2, Meta's large language model, and can generate code and natural language about code from both code and natural language prompts. This enables it to be used for code completion and debugging. In addition, Code Llama supports many popular programming languages today, including but not limited to Python, C++, Java, PHP, Typescript, C#, and Bash.

Available in three models, Code Llama offers users varying levels of functionality:

  • Code Llama: The foundational code model
  • Code Llama – Python: Specialized for Python programming
  • Code Llama – Instruct: Fine-tuned for understanding natural language instructions

How Does Code Llama Compare to Its Competitors?

Meta's Code Llama joins the ranks of AI code assistants like OpenAI's GPT-4, Amazon's CodeWhisperer, and Google's yet-to-launch Alpha Code. Trained on datasets comprised of code snippets, Meta claims that Code Llama outperforms its rivals in popular coding benchmarks such as HumanEval.

To cater to different user needs, Meta is offering Code Llama in three sizes: 7 billion parameters, 13 billion parameters, and 34 billion parameters. The lower-end model, with 7 billion parameters, is suitable for less-demanding tasks like code completion and can be run on a machine with a single GPU. In contrast, the higher-end 34-billion-parameter model boasts greater capabilities but requires more demanding hardware resources.

Facebook Wants to Kill OpenAI: A New Rivalry Emerges

By releasing Code Llama as a free-to-access AI service for writing computer code, Meta has positioned itself as a direct competitor to OpenAI's ChatGPT model released last year. This move signals an emerging rivalry between the two tech giants as they vie for control over the growing market for AI-powered tools in the coding industry.

With Code Llama's advanced features, such as code debugging and completion based on human input or prompts, it's clear that Meta aims to surpass the capabilities of ChatGPT and establish itself as a leader in the space.

The Future of AI in Coding

As artificial intelligence continues to advance at an unprecedented pace, AI tools are becoming increasingly important in various industries, including coding. With the release of Code Llama, Meta showcases the potential of AI to revolutionize the way programmers write, debug, and complete code.

The introduction of Code Llama and other AI-powered code assistants signals a new era in programming, where coders can rely on these powerful tools to streamline their workflows and improve overall efficiency. As competition grows between Meta, OpenAI, and other tech giants, we can expect even more powerful AI tools to emerge in the near future.

Share this post
Copied !
No items found.
Benoît's picture

First published on Aug 25, 2023 - Updated on Aug 25, 2023 3 min read