The prompt engineer position, which was once popular in 2023, has quickly become obsolete in just two years due to the rapid evolution of AI and the change in enterprise needs. This article provides an in-depth analysis of the three major trends behind this phenomenon, as well as the popularization and future development direction of prompt engineering skills.
The most popular AI position in the first year of the large model is now outdated –
Prompt engineers, no need to write code, are not limited to majors, do not require academic qualifications, just study how to chat with AI, and can get an annual salary of 25-330,000 US dollars in 2023.
But now, it has become one of the most reluctant positions for companies to expand.
A Microsoft survey involving 31,000 employees shows that prompt engineers have become the second to last job that the company does not want to add in the next 12-18 months.
At the same time, on the recruitment platform (Indeed), the number of searches for prompt engineers is also undergoing a roller coaster change.
- In January 2023, only 2 out of every million searches were search prompt engineers.
- In April 2023, this number skyrocketed to 144 times/million.
- Now it has returned to calm, about 20-30 times/million times.
You know, this new position was jointly recognized by OpenAI Ultraman and AI god Kapathi.
Two years have passed, and understanding prompt engineering is indeed still a skill, but the derived positions are no longer so rigid.
Executives engaged in AI course training said:
Whether you’re in finance, HR, or legal affairs, understanding prompt engineering is already a basic vocational skill, rather than needing to recruit another position.
Prompt engineering has become a basic essential skill
Sorting out the current situation, the field of prompt engineering now presents three new trends:
1. AI can automate prompt word engineering
2. The threshold for ordinary people to get started has become lower
3. Enterprises need more compound talents
Initially, prompt engineers were defined as “bringing AI to its full potential with appropriate descriptions.”
The first AI companies to open this position include Anthropic, the creator of Claude.
In March 2023, they offered an annual salary of $25-330,000 for prompt engineers, with abstract specific requirements, including hacking, liking decryption, being good at communicating, and being able to make ambiguous questions clear.
The only hard requirement is to have basic programming and QA skills, and be familiar with the architecture and operation of large models.
But two years later, there is no trace of prompt engineers in Anthropic’s recruitment.
It also launched Prompt Improver, a tool that can quickly optimize prompts.
It can automatically write or optimize prompts, which not only helps developers quickly migrate from other models to Claude, but also further improves the efficiency of enterprise-level AI development.
Coincidentally, Google also released a white paper on prompt engineering some time ago to teach non-technical users the know-how of prompts.
This means that technology manufacturers are further lowering the threshold for prompt engineering through standardized tools. Even if you are a pure novice, you can solve basic prompt engineering problems based on the above tools/solutions.
From a broader perspective, the demand for prompt engineers is generally decreasing.
A survey commissioned by Microsoft shows that prompt engineers have become one of the most reluctant positions for companies to expand (second to last), and people prefer to recruit AI trainers, AI data experts, and AI security experts.
The head of recruitment at human resources company Xpheno said that the demand for prompt engineers in the AI industry is gradually flattening, and it has been almost flat globally in the past three quarters.
The market needs compound talents who understand prompt word engineering.
For example, talents who have a better understanding of basic models, AI security governance, data and cloud computing. In the long run, the need for prompt-only engineers may gradually decrease.
But at this stage, there is still a recruitment demand for prompt engineers on the market.
In Liepin, companies that still offer high salaries to prompt engineers are more concentrated in vertical industries or technology providers serving vertical industries.
The prompt engineers they recruit also need to know more about the industry.
For example, in the recruitment of Geling Deep Pupil, it is clearly mentioned that this position requires in-depth knowledge of government affairs business.
In short, in the short term, the position of prompt engineer will not completely disappear.
But how long will it last? This may depend on the speed of AI evolution.
To achieve these three challenges, product managers will only continue to appreciate
Good product managers are very scarce, and product managers who understand users, business, and data are still in demand when they go out of the Internet. On the contrary, if you only do simple communication, inefficient execution, and shallow thinking, I am afraid that you will not be able to go through the torrent of the next 3-5 years.
View details >
In the future, everyone must understand prompt engineering
In fact, when prompting engineers to explode, some people proposed that there is an underlying logic in this position – AI is not smart enough.
At that time, the intelligence level of large models was limited, and they often needed well-designed prompts to give the answers users wanted.
Two years have passed, and the most powerful GPT-4 at that time has been removed from ChatGPT by OpenAI, and the intelligence level of large models is no longer the same, and the model can give better answers without perfect prompts.
At the same time, through some simple interaction design, large models can also understand users’ problems more accurately.
For example, the model will ask further questions to help users refine their needs.
Or, it automatically optimizes the prompt for the user’s question. For example, combine context, add details, etc.
There is also a trend that cannot be ignored, and AI is moving towards a more personalized trend.
As users communicate more and more with AI, everyone can form their own knowledge base, and AI will become more and more like a personal assistant. At this time, AI has a deeper understanding of users, and generalized prompt engineering may no longer be practical.
Finally, AI is gradually penetrating into everyone’s daily work life. Taking the domestic DeepSeek craze as an example, dialogue with AI has become no longer uncommon, and enterprises have begun to embrace the AI wave and carry out a new round of intelligent upgrades.
It is almost foreseeable that in the future, people will use AI as usual in the current use of office.
So, do you still need to spend a lot of money to hire a prompt engineer?
Obviously, a more comprehensive worker is more cost-effective (doge).
