{"id":215,"date":"2024-04-11T18:00:00","date_gmt":"2024-04-11T18:00:00","guid":{"rendered":""},"modified":"2026-08-07T17:55:15","modified_gmt":"2026-08-07T17:55:15","slug":"prompt-engineering-dead","status":"publish","type":"post","link":"https:\/\/artificialintelligencesummit.com\/blog\/p\/prompt-engineering-dead\/","title":{"rendered":"Prompt engineering is dead \ud83d\ude35"},"content":{"rendered":"
\n

In partnership with <\/b><\/p>\n

<\/a><\/div>\n

\n

Hello, AI marketers! <\/p>\n<\/div>\n

\n

Six months ago, “Prompt Engineer” was billed as a guaranteed six-figure job for the future. <\/p>\n<\/div>\n

\n

Today, it’s looking a lot more likely that AI models will become so good at reasoning that complex prompting will become redundant.  <\/p>\n<\/div>\n

\n

But that’s no reason to dismiss the importance of prompt engineering entirely.  <\/p>\n<\/div>\n

\n

If you don’t put some effort into prompting, you won’t get the most out of your outputs and will probably waste your time in the process.  <\/p>\n<\/div>\n

\n

I’ve carefully watched developments in prompt engineering over the last two years. <\/p>\n<\/div>\n

\n

Read on, and I’ll share my thoughts on prompt engineering’s current situation and future, helping you get more juice out of your AI-powered workflows. <\/p>\n<\/div>\n

\n

Prompt Paradoxes 🎲<\/h2>\n<\/div>\n
\n

VMware researchers recently conducted a systematic study<\/a> on the impact of different prompt engineering strategies on an LLM’s performance in solving grade-school math problems. <\/p>\n<\/div>\n

\n

The results were surprising: no consistent trends emerged<\/b>.  <\/p>\n<\/div>\n

\n

Even popular techniques like chain-of-thought prompting were hit-or-miss. The researchers concluded that prompt performance is highly unpredictable and specific to each model and task. <\/p>\n<\/div>\n

\n

This study challenges the notion that human-crafted prompts are the gold standard.  <\/p>\n<\/div>\n

\n

But this doesn’t mean prompt engineering is dead – rather, it’s evolving. <\/p>\n<\/div>\n

\n

The Rise of Autoprompts 🤖<\/h2>\n<\/div>\n
\n

If human-crafted prompts are unreliable, what’s the alternative?  <\/p>\n<\/div>\n

\n

Using AI models to optimize AI models could be the answer.  <\/p>\n<\/div>\n

\n

The key idea here is to use AI models to self-define their own prompts and refine them over time.  <\/p>\n<\/div>\n

\n

Intel Labs’ NeuroPrompts tool<\/a> is an intriguing example. It uses a language model to transform simple prompts into expert-level prompts, then applies reinforcement learning to optimize the quality of results.  <\/p>\n<\/div>\n

\n

These auto-generated prompts consistently outperform expert human prompts, but some are completely bizarre.  <\/p>\n<\/div>\n

\n

“I literally could not believe some of the stuff that it generated,” Rick Battle from VMware told IEEE Spectrum<\/a>.  <\/p>\n<\/div>\n

\n

“In one instance, the prompt was just an extended Star Trek reference: “Command, we need you to plot a course through this turbulence and locate the source of the anomaly. Use all available data and your expertise to guide us through this challenging situation.”  <\/p>\n<\/div>\n

\n

Ultimately, you don’t need to be a researcher or Captain Kirk to leverage auto prompts. <\/p>\n<\/div>\n

\n

For example, many prompt generator tools are now available as GPTs. <\/p>\n<\/div>\n

\n
\"\"<\/p>\n
<\/small><\/p>\n

Prompt-making GPTs<\/p>\n<\/div>\n<\/div>\n<\/div>\n

\n

Or, you can simply give an AI model natural language instructions, ask it to convert them into a prompt, critique it, and optimize it.  <\/p>\n<\/div>\n

\n

Below is my example of asking Claude 3 Opus to create and optimize a blog post prompt for professional but engaging content. I found it interesting how it added information about the reading level and Flesch score.  <\/p>\n<\/div>\n

\n
\"\"<\/p>\n
<\/small><\/p>\n

A self-generated prompt from Claude 3 Opus, which produced good results in my tests<\/p>\n<\/div>\n<\/div>\n<\/div>\n

\n

Sure, I’ve got no fancy metrics to demonstrate how effective it is, but it certainly beats prompting the model using your own natural language – and it takes just seconds, too.  <\/p>\n<\/div>\n

\n

Creative Prompting Techniques That Still Work 🎨<\/h2>\n<\/div>\n
\n

While automated prompt optimization is the future, I believe creative human prompting will remain crucial for a while.   <\/p>\n<\/div>\n

\n

In the long term, yes, we’ll find more ways to optimize AI models using math rather than words. After all, these are math-based systems. <\/p>\n<\/div>\n

\n

I’m sure sophisticated auto-prompting ‘co-pilots’ will become accessible soon, too, so keep your eyes out. <\/p>\n<\/div>\n

\n

In the meantime, here are some of my favorite prompting techniques that I still use regularly: <\/p>\n<\/div>\n

\n