Does Prompting Still Matter? The New Rules for Getting What You Want From AI

The Old Prompt Engineering Playbook Has Faded

Remember when getting a great answer from ChatGPT meant knowing the right magic words?

You had to tell the model to act like an expert, think through a problem step by step and follow a carefully constructed set of instructions. Entire guides were devoted to prompt formulas, elaborate templates and tricks for squeezing better performance out of language models.

We even published our own guide to role-based prompting in March 2025. The advice reflected a widespread belief: assigning the model an expert persona could improve the quality of its answers.

Fast-forward to today, and Forward Future producer Alex McCraw doesn't bother with that technique anymore.

“I used to use role-based prompts like ‘You're an expert graphic designer,’ with the idea that assigning the AI an expert role would make it perform better,” Alex says. “I don't really bother with that anymore.”

He's not alone.

After some internal debate arose, we polled our team to see whether prompt engineering is still essential. For such an opinionated group, our responses were surprisingly aligned.

Forward Future founder Matthew Berman, the person behind our AI model reviews, believes traditional prompt engineering is essentially finished.

“Prompt engineering, which I define as extracting the best performance from a model with techniques (old) such as ‘think through this step by step’ is basically useless now,” he says.

Our editor Brian Knepprath goes even further: “Over-specifying or spending too long on a prompt is now actively a waste of time.”

Forward Future co-founder Nick Wentz agrees that the old techniques are losing some relevance, but he thinks something else is happening at the same time: The bar for great prompting is getting higher.

“The main shift is from telling AI how to do something to communicating what you want, why it matters and what good looks like,” he says.

In Nick's view, the best AI users aren't necessarily the ones with the most sophisticated prompts. They're the ones who understand how to frame a problem, provide relevant context and recognize a good result when they see one.

That's a remarkable consensus from people who spend their working lives experimenting with AI. But there's a catch.

The same people who believe prompt engineering is dying also consistently produce results that less experienced users struggle to replicate. In other words, our poll respondents aren’t representative of the broad range of people trying to get work done with AI.

So what's the difference?

We think the answer is that the better AI gets, the less you need to tell it how to think. But — you still need to know what you want, which instructions are critical and when to get out of its way.

Let's start with the techniques our team has abandoned.

Matthew points to the instruction “think step by step,” once a common way to encourage language models to work through complicated problems. Alex has similarly abandoned expert-role prompts and attempts to pressure models into performing better.

With the arrival of reasoning models, that kind of instruction has become much less necessary because the models are trained to reason through complex tasks without needing users to explicitly instruct them to do so.

Brian's example is more recent. Until just a few months ago, he frequently used /goal, a command intended to help agents stay focused on a particular objective.

Now he rarely uses it.

“I've been finding more and more that regular prompts accomplish what you ask for, whether you're asking for a lot of stuff or a little stuff, and whether you have goal in it or not,” Brian says.

This is reflected in official guidance for modern reasoning models. OpenAI has advised developers that simple, direct prompts often work best and that explicitly requesting step-by-step reasoning can be unnecessary or even counterproductive.

Nick has also abandoned elaborate personas. He used to ask models to write like respected authors or approach problems as world-renowned researchers. Today, he finds those instructions rarely improve the results.

Matthew offers a revealing example from his software-building experiments. When he asks AI to create something resembling an existing product, such as an Excel-style application, he no longer needs to describe every familiar feature in painstaking detail and precisely what he wants to use it for.

Models now understand what a spreadsheet application is supposed to do, and they know what Excel’s interface looks like. That knowledge allows Matthew to focus on the result rather than explaining the implementation.

So, if a short request can produce a convincing application prototype, how much time should anyone spend perfecting the prompt? Not a lot.

Here's where the debate gets interesting.

Alex and Brian agree that elaborate prompting is becoming less necessary, but they have different approaches to specificity.

Alex believes detailed instructions remain essential when you have a particular outcome in mind.