The continuous advancement of artificial intelligence in more and more areas one of them has to do with the programming code, in which Alphabet (Google’s parent company) is working in a way that will develop an artificial intelligence capable of both creating code ex novo how to correct another code already written earlier.
Artificial intelligence replicates processes of the human intellect through algorithms and instructions, becoming able to learn on its own
This project is part of a larger project in which artificial intelligence is able to use algorithms to create images, videos… and also code. This project, known by the code name Pitchfork (Spanish for “pitchfork”, in reference to the tillage tool) went this summer to Google Labs, the special projects development department.which would indicate the importance attached to it within Alphabet.
The Google Labs department has been in charge, over time, of developing both leading projects that have become final products and services (such as Google Glass) but also discarding others that failed to have a practical or usable utility.
Specifically within Google Labs the development of Pitchfork falls on the department called Artificial Intelligence Development Assistance, led by Olivia Hatalsky, an employee with a long career in Google X and who already worked, among other notable projects, in the development of Google Glass.
In the case of artificial intelligence this year 2022 has undoubtedly been the one that has surprised with those that allowed to generate realistic photo images from text descriptions with names like DALL-E, Midjourney or Stable Diffusion.
Pitchfork has been developed with the aim of being able to learn to write code and even be able to rewrite its own code, with the intention of applying what it “learns”, given that artificial intelligences are distinguished by being able to emulate reasoning processes similar to humans but also to learn on their own, applying the acquired knowledge.
In this case the work is focused on getting the research team to be able to find applications and ways in which artificial intelligence can help developers.