A quick recap of recent AI-related booming technologies & related questions
Recent months have seen a surge in AI-related technologies, such as those related to machine learning, natural language processing, and computer vision. This quick recap provides an overview of these developments for those who may have missed them.
Quick Recap
I do not intend to provide an exhaustive list of AI initiatives, but simply some major trends driven by certain projects with a global impact.
DALL-E and other image-related technologies
Last year, the world was introduced to DALL-E (OpenAI), an AI system that could change how we use technology. DALL-E showed its strength by creating complex pictures from short descriptions. This system had immense potential, which could create new possibilities in AI, computer vision, and natural language processing. It was a thrilling moment since machines could do more and more.
I published a complete article last year dedicated to AI-generated art.
ChatGPT
Recently, ChatGPT came out and it has a powerful dialogue engine that can answer almost any question you ask. It gives answers that are very detailed and can understand how people talk. ChatGPT can understand complex orders and give answers that sound almost like a real person.
Once again, this tool comes from OpenAI’s APIs. It is OpenAI that provides the APIs of some well-known services to developers, such as GitHub’s Copilot, or Notion’s AI.
VALL-E
Have you heard of “Deep Fake”? This technology allows you to take the face and voice of a person and then make them say what you want.
VALL-E is positioned on the voice and the results produced and available are absolutely stunning in realism.
I really encourage you to visit VALL-E GitHub’s page and see how it works, and how powerful it is.
What to do with these technologies?
Once the initial fun tests have been passed, we quickly realize the power of the tools.
AI for personal organization
Fun for testing purpose but useless according to me :)
AI for learning
You ask almost everything and you can add some context parameters. Useful for “lazy” learning and learners.
AI for producing
I did several tests, including the creation of functions, or complete application creation. It’s working quite well as far as we manage to provide proper details to the bot. Here I’m sharing an example of a simple regex function, but I managed to develop a functionally working Google Chrome plugin for example.
AI for improvement
This is super powerful. I voluntarily pushed a piece of code including a typo, and he detected and corrected it.
What’s next & personal analysis
We must embrace these new capacities
I sincerely believe that we must now live with these new capabilities and make the best use of them.
Risks
As with any technology, there will be good and bad, or at least it will be necessary to position the cursor on what is acceptable or not.
Will people stop learning and thinking? How to judge the quality of a job when there is doubt of an AI? How will security teams in companies react? There are plenty of questions and new challenges that come with the use of these technologies.
How will schools and teachers react? In my opinion, blocking these technologies would be counter-productive. On the contrary, it is necessary to make students aware of the possibilities and limits, and above all to provide them with a critical view of these tools. Just as the arrival of calculators in school caused debate many years ago, it is highly likely that AI will be part of the “normal” landscape of students in the years to come.
Another highly sensitive area is that of security. How can we be sure that we are talking to the right person when it is technically possible to reproduce a voice, a face or a writing style?
Opportunities
AI is everywhere, I don’t see any area which won’t be impacted by these technologies: from automotive to healthcare, from creation to final products, from arts to manufactures…
From a professional point of view, I think these technologies can bring a lot of improvements in quality, optimization, and time to market.
Some examples that come to mind:
- I was talking in another article about the development of testing strategies which, in my opinion, can only be done with AI. Technology can currently produce Gherkin with no effort!
- Increase code quality, optimize code.
- Accelerate developments by automating recurring tasks.
- Produce code based on good specifications.
- Create and deploy automatically new cloud environments.
- …
As for me, I am extremely excited about these new technologies, this new field of possibilities, and I can’t wait to see what the next few months will bring as technology is developing extremely quickly.
I strongly recommend that you dive into OpenAI’s research articles, it’s exciting.
