how Prompt Engineering is Changing Technology in 2023
Even with the constant improvements to model generation, there will always be a need for someone to prompt it accordingly to output optimal results.
In this article, we cover the current developments in both generative AI and prompt engineering which is the main tool right now to interact with text-to-output models.
background on Prompt Engineering
In layman’s terms, AI prompt engineering is the act of crafting a prompt to extract a more desired output from the LLM.
The term came about along with the emergence of large language models such as ChatGPT.
Despite it garnering major attention in the tech sector, widespread awareness of what it is and what it entails remains quite low. At least from what I have observed. It is growing though!
an Example?
A common use of prompt engineering is role assignment to illicit a specific type of results.
For instance, if you were seeking detailed physical therapy advice due to a sports related injury you sustained, you could prompt the following:
prompt: “Imagine you’re a seasoned sports physiotherapist, and a dedicated athlete seeks your expertise to recover from an recent injury. Provide a comprehensive physical therapy plan given the injury is [input injury details here]. Cover exercises, duration, frequency, and any necessary equipment. Keep the advice professional and supportive, offering encouragement and expert guidance throughout.”
This prompt will provide you with a comprehensive plan of action to help deal with your injury.
ChatGPT output from prompt engineer example = ‘muscle soreness’
Whilst you could simply prompt ChatGPT with your injury, the response would be more general advice with no plan for progression. Try writing ‘I have muscle soreness in my left arm’ and compare the results.
General prompting like this is a thing of the past but new AI tools, models and frameworks are bending a knee to this type of prompting so that even casuals can obtain more useful results.
the Latest in Generative AI
DALL-E 3
Open AI’s DALL-E 3
DALL-E 3 offers text processing improvement over its counterpart, DALL-E 2.
It now takes your the input prompt and builds on it using ChatGPT which is similar to asking ChatGPT to generate an image description given an input theme. This means the user input can be a lot more vague and still generate great results.
Plus DALL-E 3 is directly integrated into ChatGPT Plus making it easier than ever to craft the perfect image.
Upon first hearing about DALL-E 3, many suggested that this was the end of prompt engineering but this certainly not the case as you still need to interact with ChatGPT which of course requires a prompt still.
This release is among an avalanche of generative AI developments occurring daily.
chatbots, chatbots and more chatbots
chatbot with multiple functions
The uses for chatbots are being discovered for many industries such as medicine and health, scientific research and more.
They are built on an LLM such as Llama or ChatGPT and then are engineered to output more tailored responses. Guess how? Through prompt engineering.
An interesting chatbot I found out about last week was “Texts from my ex” which basically reads over your conversations with your partners to optimise your dating experience for the better. It is GPT-powered and gaining quite a bit of traction.
increasing Investment
Visa are amongst the huge pool of companies to hedge their bets on generative AI due to its wide applicability and ability to boost customer service quality. The firm apparently is eyeing $100 million to invest into generative AI companies.
To add to that, Amazon recently decided to partner with Anthropic, an AI research and safety company to help give their AI plans a sense of ethical yet innovative direction. They pledged $4 billion to their cause!
merger between Anthropic and Amazon
Anthropic specialise in prompt engineering which is a major tool for making useful chatbots. The partnership with Amazon is likely to bolster their customer service through use of chatbots and other AI tools once again emphasising the cruciality of prompt engineering.
Those are just a few public announcements. The amount of interest and attention generative AI has gained in the past year is remarkable and a good indicator of things to come.
the Future of Prompt Engineering
The skill started out as a way to generate more useful answers from ChatGPT but whole industries are now being revolutionised by the skill.
I am referring to increasing widespread interest in generative AI and in particular chatbots which are optimised using prompt engineering techniques.
This interest is being reflected in some of the salaries being reported for prompt engineers. Some upwards of $300,000 but since it is a relatively new as job, salaries are still largely unknown.
Based on the information I found, the future of prompt engineering is a bit up in the air right now, but one thing is for certain: prompt engineering will continue to play a vital role in creating efficient communication between humans and machines.
what Are We Doing at Asycd
Here at Asycd, we recognise the need to adapt with the times and prepare for the revolution which is generative AI.
With that said, we are pleased to say we are working on our first chatbot which will fine tuned to your prompts before you input them into an LLM.
We want to give people the tools to build their own future which is why we think this solution is needed.
The main architecture will be based on the function-based prompting.
Thanks for reading. If you are looking for more thought-provoking articles about technology, art and more, check out our blog page.
