Startup specializing in AI predicts that relying on humans to write programs will be outdated in the future And reignites the debate about the possible demise of the developer profession

When we talk about artificial intelligence, two main currents of thought collide: that of third parties who think it is a tool, and that of stakeholders who believe that it is only a matter of time before it becomes a threat to humans. Matt Welsh, from startup Fixie.ai, suggests that he is in the latter group when he says: Relying on humans to write programs will become outdated in the future. However, critics believe that AI does not pose a threat to software developers, but an opportunity to reflect on the evolution of this role in the coming years.

Matt Welsh portrays a future in which the IT industry will move from the approach of writing computer programs by humans to the contribution of previously trained artificial intelligence agents to replace humans in activities related to the developer profession. The latter is of the opinion that artificial intelligence has the shoulders to take on this change in view of the rapid progress of the sector:

Programming wizards like CoPilot only scratch the surface of what I’m describing. It seems pretty obvious to me that, in the future, all programs will be written by artificial intelligences, with humans relegated, at best, to a supervisory role. Anyone who doubts this prediction need only look at the very rapid progress being made in other aspects of AI content generation, such as imaging. The difference in quality and complexity between DALL-E v1 and DALL-E v2 – announced just 15 months later – is staggering. If I’ve learned anything over the past few years working in AI, it’s that it’s all too easy to underestimate the power of ever-greater AI models. Things that seemed like science fiction just a few months ago are quickly becoming reality.

I’m not just talking about Github’s CoPilot replacing programmers. I’m talking about replacing the very concept of writing programs with dedicated artificial intelligence agents. In the future, computer science students won’t need to learn mundane skills like adding a node to a binary tree or coding in C++. This type of teaching will be outdated, like teaching engineering students to use a slide rule.

The latter goes even further by adding that the use of artificial intelligence as the basis of computing is likely to open up the sector to everyone. You no longer need to be a computer programming expert to get something useful from an artificial intelligence, he says, thus reviving the prospect of the demise of the developer profession.

Stakeholders in the IT development sector are, however, of the opinion that artificial intelligence is only good for automating redundant tasks and will not go beyond that level. The development of specifications or specifications for large IT projects continues to be a task that demands a creativity of a human brain that an artificial intelligence will never have.

The biggest debate revolves around the possibility of seeing machines returned to the stage where they are endowed with common sense, capable of causal reflection, that is, this ability to reason about why things happen.

Google, for example, secretly launches into Pitchfork development, or AI Developer Assistance. It’s a tool that uses machine learning to teach code to write and rewrite itself. How much? Learning corresponding styles of programming languages ​​and applying that knowledge to writing new lines of code.

The original intention behind this project was to create a platform capable of automatically updating the Python codebase every time a new version was released, without requiring intervention or hiring a large number of engineers. However, the program’s potential turned out to be much greater than expected. From now on, the intention is to give life to a general-purpose system capable of maintaining a quality standard in the code, but without depending on human intervention in development and update tasks. Such a goal could no longer be science fiction when we know that research teams in artificial intelligence already promise to reach the stage of general artificial intelligence in 5 to 10 years.

source: cacm

Is that you?

Do current developments in the software engineering pipeline give rise to legitimate concerns about the future of human computer scientists in the pipeline?
What does the possibility of research leading to artificial general intelligence suggest to you in 5 to 10 years?
How do you see artificial intelligence in 5 10 years? As a tool or as a danger to your work as a developer?

See too:

Is autonomous driving today just a futuristic vision of Tesla Motors? The company just changed the goals of its autopilot
SEC asks Musk to step down as Tesla chairman, demands $40 million fine for out-of-court settlement
Tesla announces that the new computer for fully autonomous driving of its vehicles is in production and will be tested this month
Tesla shares tumble after its Autopilot system is involved in an accident and reports of vehicle batteries catching fire

Leave a Comment