As more and more enterprises try to use AI basic models (such as OpenAI’s ChatGPT), but at the same time, the impact of automation platforms on software developers is increasing obvious.
On the one hand, these platforms may revolutionize the way developers work; on the other hand, these platforms also threaten the jobs of those in the industry and stand in stark contrast to established processes.
That said, there is no doubt that platforms like ChatGPT are a disruptive force and have been enthusiastically welcomed by venture capitalists. According to data from research firm Gartner, generative AI solutions have received more than $1.7 billion in investment in the past three years, a large portion of which has been invested in AI software coding.
Automation has many other potential impacts on software development, not least the quality and diversity of generated code.
Industry analyst Dave Vellante writes: “Large language models can reduce complexity and accelerate the adoption of enterprise automation platforms, but the flip side is that software robots are designed to increase human productivity through intelligent automation, and the GPT model "It may cannibalize some, if not many, of the use cases initially targeted by RPA vendors."
This new reality is prompting customers to rethink their automation strategies and vendors to refine the messaging they deliver. , positioning the base model as the catalyst for their platform.
So how will this new technology change the open source developer community? Why are emerging automation platforms a double-edged sword? Finally, how will this change the industry forever?
Before the explosive development of AI, automation had already reached an alarming rate of development. According to recent forecasts by the World Economic Forum, automation and technological advances are expected to displace 85 million jobs by 2025, but create 97 million new roles "as humans, machines and algorithms increasingly work together."
Open source generative AI, such as the natural language processing model GPT-3 developed by OpenAI, may subvert most technical fields, and software developers are not immune.
For example, GPT-3 can automatically generate code, and some startups, such as Hugging Face, have begun to prepare for this.
Hugging Face CEO Clement Delague said: "Hugging Face is a resource for startups and other enterprises around the world, and we can help them build almost any natural language processing application without the time, cost and complexity. They develop a fraction of what they use themselves, helping organizations bring their solutions to market quickly."
We are testing internally," GitHub CEO Thomas Dohmke said in a recent blog post. With the new feature, GitHub Copilot will automatically make suggestions for sentences and paragraphs when developers create pull requests by dynamically extracting information about code changes."
AI technology is already being used to optimize code and propose improvements Suggest and reduce the possibility of errors, in addition to being used for automated testing. Although ChatGPT has achieved a "Netscape-like" moment in the technology world, it's unclear how IT decision-makers will directly implement this technology in their organizations.
In an analysis, Vellante cited the contents of a recent CIO roundtable. When talking about robotic process automation, machine learning and artificial intelligence, one CIO said that all three are essentially trying to solve the same business scenario, which is to eliminate redundant resources, whether human resources or other resources.
The CIO said: "OpenAI showed with ChatGPT that you can avoid a lot of the 'overhead' of building complex artifacts around typical RPA. I think that is a very interesting value proposition - being able to replace ' There are some workbenches in classic 'RPA that take a long time to master and a long time to get any value beyond regular use cases."
Others are trying to make these technologies more accessible . Red Hat is committed to creating community-driven AI as part of its Project Wisdom initiative to simplify the automation of infrastructure.
Tom Anderson, vice president of Red Hat and general manager of the Ansible business unit, said: "Now this is the beginning of the community journey, and we will collaborate on this work through channels like Discord to attract people to participate."
The other side of the issue involves the potential for this emerging technology to completely replace human developers. In fact, companies such as Goldman Sachs Group Inc. are already experimenting with using generative AI internally to assist in writing code.
If an organization relies too much on AI-generated code, other problems will arise. For example, if there are different basic models for different tasks, fragmentation will occur, which may make it difficult for developers to collaborate effectively. In addition, problems may arise if there are no checks and balances on the code generated by AI, especially considering that so-called "hallucinations" may still occur in ChatGPT.
Neuroflash co-founder and chief marketing officer Henrik Roth said: "These technologies will very convincingly tell you what it 'thinks' is right, no matter how many errors it has." Roth said that ChatGPT is used A great tool for creative writing and advertising, but “in areas where facts are more important – such as journalism and science – one should fact-check every claim.”
Automated platforms will inevitably change the rules of the game for software developers. Large enterprises, especially those at the intersection of machine learning, artificial intelligence and automation, are using RPA as an alternative tool, including Microsoft, AWS and Google.
Vellante said: "Microsoft is well-positioned in terms of revenue share, and observers can expect that Microsoft's relationship with OpenAI will become key to Microsoft's AI strategy. From a technology perspective, Microsoft can be said to lag behind in AI. AWS and Google, but in terms of business models, Microsoft seems to have jumped into the lead."
In his research on enterprise technology, customers were asked whether they were evaluating the GPT model and for which use cases. A surprising 56% of customers said they were not evaluating this, and a majority were evaluating the technology for customer chat.
"On the surface one would conclude that RPA and automation platforms can benefit from the GPT model and that these use cases are largely complementary. For example, the base model can be used to write code or accelerate guidance Automated code development for software robots, but at the same time, there is overlap between the capabilities of large language models and some early RPA use cases, and this overlap is likely to increase further over time."
For developers, opportunities and challenges loom, as underlying AI models have the potential to revolutionize the way they work, even as threats to job security and established processes remain.
The market remains polarized: In tough times, companies look for ways to cut costs and look for areas where automation technology can make a difference. But at the same time, businesses have to spend money to make money.
"The GPT model is spawning new ideas, and buyers and sellers are working hard to turn the underlying model into opportunities. The early use cases of the GPT model are interesting, but they are not a direct replacement for enterprise automation platforms. However, low-end automation There are risks, and there's no doubt there's a Venn diagram here intersecting the underlying model and the automation platform."
In fact, generative AI will cannibalize some RPA use cases. On the other hand, the two technologies can work together to automate a wider range of tasks.
“Nonetheless, we believe all vendors must leverage the GPT model to simplify and accelerate adoption; buyers must take a step back, test the waters, and see how they deploy these new innovations for their businesses Add value."
In the same way that AI advancements have disrupted most technologies, automation platforms are constantly changing, and only time will tell what the future will look like.
The above is the detailed content of How will artificial intelligence change the world of software developers?. For more information, please follow other related articles on the PHP Chinese website!
![[Web front-end] Node.js quick start](https://img.php.cn/upload/course/000/000/067/662b5d34ba7c0227.png)
