The artificial programming wars begin

A group of 2,000 Artificial Intelligence (AI) experts have signed a letter calling for a halt to all AI experiments and urging governments to review current practices to better regulate these advancements. They argue that we are not dealing with intelligent machines, but rather with probabilistic parrots. These machines constantly replicate what they learn without truly understanding what they are processing. They are not conscious machines.

However, what I do think is worrying is the accelerated job displacement that seems to be coming. Perhaps that's why they're asking for time to understand the misinformation, as well as the social and labor impact this will have. One of the professional fields where the most transformation is expected is in the area of ​​programming. And asking one of these machines for source code is easier than ever. Our artisanal work of producing a computer program has given way to curation, like that of media editors and those who curate works of art in museums. We monitor what our co-pilot has done. Proof of this is the following: we asked ChatGPT to create a timer with a specific behavior. It understood, and it tells us how to deploy the resulting program:

Even if OpenAI, Google, Microsoft, and other tech giants were to stop doing what they're doing—and they're not going to—the AI ​​models they've already developed are likely to have profound repercussions, especially in software development. Alphabet's agreement to supply AI to Replit , a web-based coding tool with over 20 million users, is something of a seismic shift. Replit will use Google's AI models, along with others, in Ghostwriter , a tool that recommends code and answers code-related questions in a way similar to ChatGPT. Google will also make Replit available to Google Cloud users, helping it reach more enterprise customers.

The move is especially significant because Alphabet is going up against Microsoft and GitHub, which also use AI to assist programmers with Copilot . When you start writing code, tools like Copilot suggest ways to complete it. Alphabet's move also signals what could be the next major battleground for Big Tech. While so much attention is focused on ChatGPT and Donald Trump's MidJourney 5 builds , the big story is which company can offer developers the best AI tools and the new software they'll create with that AI by their side.

Microsoft's research suggests that programmers can now perform tasks 50% faster with these AI assistants. Thus, companies offering AI as programming co-pilots can attract developers to their coding tools and get those users hooked on their cloud computing services. Amazon has developed an AI programming tool called Code Whisperer, and Meta is also working on one for internal use. Presumably, Apple won't want to be left behind.

In addition to helping developers write code, AI is beginning to change how code is composed. Last week, OpenAI announced the creation of the first plugins for ChatGPT. Thanks to these, the bot will be able to perform tasks such as searching for flights, booking restaurants, and ordering food. Incorporating AI into code can also accelerate software development. This week, Masad from Replit shared a good example: an application that will turn voice commands into functional websites. With the speed at which things are moving, it's worth considering what consequences the rapid integration of AI into software development might have. AI tools can reproduce vulnerabilities in code that developers may not notice or be able to detect.

Perhaps developers will become more complacent, or see their skills atrophy, if they rely too heavily on AI. And what kind of " technical debt " might arise if programmers have to go back and fix software that no human has ever closely examined?

In short, soon, when we seek to write software code in any programming language, we will use copilots from some of the aforementioned artificial intelligences. In fact, the name of this last tool is how I envision the future human-machine coexistence: copilots (technologies) accompanying pilots (humans). This is because the probabilistic parrot behind these AIs is capable of reproducing all structured and formal language, just like programming in the vast majority of programming languages.

Brain and Code ©
April 8, 2023