The writer tears down prompt engineering while extolling the virtues of well-written prompts. Also GPT-4 isn't "learning new things all the time" in that way...it's getting updates, sure, but not constantly or in real time. These articles are pure nonsense.
So long “prompt engineering,” we hardly knew ya
11–20 of 76 posts
Re: So long “prompt engineering,” we hardly knew ya
#12i feel like people that get really into prompt engineering seem to forget that other people are working on building models that are even better . Do they really think that the tricks they developed to make gpt3.5 work well will still be necessary on gpt9 (or whatever new model replaces gpt)? if you want a long-lasting career in AI you need to work on the actual AI stuff, not just using the AI stuff.
If you want a long lasting career in programming, work on building the languages, not just using the languages.
Is probably what someone said many decades ago.
Re: So long “prompt engineering,” we hardly knew ya
#13i feel like people that get really into prompt engineering seem to forget that other people are working on building models that are even better . Do they really think that the tricks they developed to make gpt3.5 work well will still be necessary on gpt9 (or whatever new model replaces gpt)? if you want a long-lasting career in AI you need to work on the actual AI stuff, not just using the AI stuff.
If you want a long lasting career in programming, work on building the languages, not just using the languages. Is probably what someone said many decades ago.
And in those decades they were probably right.
Re: So long “prompt engineering,” we hardly knew ya
#14everyone thinking Prompt Engineering will go away dont understand how close Prompt Engineering is to management or executive communications. until BCI is perfect, we'll never be done trying to serialize our intent into text for others to consume, whether AI or human.
Where I DO think Prompt Engineering goes away is because codegen and code orchestration of LLMs rises up to take its place. Hence Prompt Engineer -> AI Engineer https://www.latent.space/i/131896365/the-role-of-code-in-the...
Re: So long “prompt engineering,” we hardly knew ya
#15i feel like people that get really into prompt engineering seem to forget that other people are working on building models that are even better . Do they really think that the tricks they developed to make gpt3.5 work well will still be necessary on gpt9 (or whatever new model replaces gpt)? if you want a long-lasting career in AI you need to work on the actual AI stuff, not just using the AI stuff.
I'd lean yes -- what data set can they feed into the training algorithm to replace the stuff that doesn't get into writing?
Re: So long “prompt engineering,” we hardly knew ya
#16I hate approximately everything about this article, but I'm glad that I took a second look through because this is a decent framework for the-thing-which-he-swears-isnt-prompt-engineering: --- * diagnosis * decomposition * reframing * constraint design. Diagnosis is discovering the problem that AI can solve. This is the human part of knowing that a problem exists. Learning to ask the right questions, look at the diff…
That's just systems analysis though. To call rediscovering that prompt engineering just because you're dealing with AI rather than programmers siloed away from business processes feels a little silly and just being justified to keep the title when as originally formulated by the people who first started using it, it was about learning the "magic words" for a specific LLM
See I always saw prompt engineering as the ability to author a prompt that elicits a specific behavior by the LLM, not simply the incantation to make to it say penis or talk like a sailor. The “make the AI be lewd” was a parlor trick, but the art of making it behave the way you want and to get the answers you want at the depth, breadth, and form you need is a skill and one I think isn’t going away in our lifetime. I’d further note that writing to influence humans to your intended response and understanding is a widely studied field by every study with multiple PhD and beyond disciplines associated. We even elect such people to lead us largely through their ability to prompt engineer humans.
IMO - Folks seem to be reacting to the term “engineer” in the same way traditional engineers with engineering exams do to software engineers claiming to be engineers despite often graduating from a LAS school or not at all - I have a bunch of PE EE in my family and they treat my claim to be an engineer with extreme scorn.
Re: So long “prompt engineering,” we hardly knew ya
#17As someone who’s spent an unreasonable amount of time working with prompting for both text and text-to-image, I can confidently say prompt engineering does work for any nontrivial use cases (e.g generation with constraints). I don’t like that prompt engineering works, but it does, and proper use of it does result in objective improvements in generation quality.
The latest ChatGPT model, gpt-3.5-turbo-0613 has better system prompt steerability, and with some prompt engineering I can get GPT-4 quality results out of it at a fraction of the cost.
Ultrapopular tools like LangChain and AutoGPT are essentially just prompt engineering under the hood.
Re: So long “prompt engineering,” we hardly knew ya
#18i feel like people that get really into prompt engineering seem to forget that other people are working on building models that are even better . Do they really think that the tricks they developed to make gpt3.5 work well will still be necessary on gpt9 (or whatever new model replaces gpt)? if you want a long-lasting career in AI you need to work on the actual AI stuff, not just using the AI stuff.
The culture war surrounding prompt engineering is so dumb. My eyes roll hard into the back of my head every time I read a comment like this.
Prompt engineering is not the same as programming. In some regards, it is “better”. In other regards, it is “worse”. They’re two different yet similar disciplines, each with their own strengths and weaknesses, but both equally legitimate.
Programming is simply the skill of being able to communicate with a compiler/interpreter effectively.
Prompt engineering is simply the skill of being able to communicate with an LLM effectively.
I wish we weren’t so divided over an incredible new technology.
Re: So long “prompt engineering,” we hardly knew ya
#19Even if these models become AGI someday, there's still sort of "prompt engineering", because there is still social engineering. What and how you say something is always going to have an effect on the response.
Re: So long “prompt engineering,” we hardly knew ya
#20i feel like people that get really into prompt engineering seem to forget that other people are working on building models that are even better . Do they really think that the tricks they developed to make gpt3.5 work well will still be necessary on gpt9 (or whatever new model replaces gpt)? if you want a long-lasting career in AI you need to work on the actual AI stuff, not just using the AI stuff.
If you want a long lasting career in programming, work on building the languages, not just using the languages. Is probably what someone said many decades ago.
Is that not true to some extent? In the early days of computer science, when there were so many new developments, it would serve you well to be the person working on writing the programs and building computer hardware, rather than the person using computers to get stuff done. The person using computers was still miles ahead of the person not using computers but where you really wanted to be was building languages and inventing new tools.