transcribe

Why The Prompt Matters Less Than The Context

Hamel Husain · 5m · transcribed Sep 2026
More from Hamel Husain Business
𝕏 Share ▶ YouTube 📥 PDF 🤖 .md

Section Insights

# 0:00

Evaluating Claude's Writing Style

What are the strengths and weaknesses of Claude's writing style?

Claude's writing has improved with fewer stock phrases and less jargon, but it can still be dense and difficult for readers. The use of mannered prose can obscure meaning and make reading harder.

  • Claude's writing is a step up but still has issues.
  • Mannered prose can irritate readers by making them work harder.
  • Direct statements are preferred over metaphorical language.
# 1:01

The Impact of Sloppy Writing on AI Output

How does the inclusion of sloppy writing in prompts affect AI output?

The expectation is that including sloppy writing in prompts would lead to more sloppy output, but initial findings suggest that it might not be the case.

  • Dropping slop into prompts challenges conventional expectations.
  • There is curiosity about how context influences AI writing.
  • The effectiveness of prompts may vary unexpectedly.
# 2:02

Comparing Different Writing Prompts

What are the results of comparing unmannered, plain writing with other prompts?

The comparison reveals that there isn't a clear winner among the different writing styles, with each having its strengths and weaknesses.

  • No single writing style consistently outperforms others.
  • Unmannered prose showed surprising effectiveness.
  • Different prompts can yield varying results in clarity and engagement.
# 3:03

The Complexity of Writing Prompts

What conclusions can be drawn about the effectiveness of writing prompts?

The analysis indicates that the effectiveness of writing prompts is not straightforward, and there may be valuable elements in each style.

  • Writing prompts do not guarantee superior results.
  • There is potential value in combining elements from different prompts.
  • The context of the prompt may play a significant role in the outcome.
# 4:05

Reassessing the Importance of Prompts

Do prompts really matter in AI writing?

The findings suggest that the effectiveness of prompts may not be as critical as previously thought, with context potentially being more influential.

  • The role of prompts in AI writing is more complex than assumed.
  • Context may be the primary driver of writing quality.
  • Surprising results challenge the belief that handcrafted prompts are always superior.

Transcript

0:00 We have this prompt guidance. Claude's fables 5.1 writing is generally a step up from earlier claude models cuz like writing is the most painful part about using AI. I think we can all agree with fewer stock phrases and less unexplained jargon. In some cases though its pros is denser than Claude's fables five sentences run longer blah blah blah and instruction that defines the time pattern manner pros helps. And it says like use this, right? And so if you look at the prompt, if you read the prompt, it says like mannered pros substitutes metaphor and flourish for a direct statement. Instead of a parameter worth varying, the mannered writer produces a dial worth turning.

0:41 and honestly, the funny thing is is that like both of those, a parameter worth varying or a dial worth turning, both of those are bad in my opinion. The phrases exist to display the writer, not to convey the idea. And readers can tell that sentence seems okay. I guess that is why mannered pros irritates. Oh my god. It makes the reader work harder so the writer can perform. I mean, that is pure slo. Let's be honest. And then it is also imprecise. Metaphors drag in connotations the writer did not choose and cannot control. The fix is here we go again. to say what you mean when a literal phrase is available use it. And they said like this short version also tends to work which is fascinating. But I'm like when I saw this I'm like wow okay why would dropping slop into a system prompt work? Like it defies my intuition like kind of like in context examples. Like if you have an incontext example that's slop you would think that that's going to sort of steer the model towards more slop. If you have slop, even if the slop is saying stop writing slop, you would think it's like not the best thing. So, I think this is what we we want to find out right now.

1:55 >> Yeah. I would think that whatever is in its context, it'll create more of that. I hope this doesn't work because I don't want to always be writing by props like that. Like I don't I don't want like the future of prompt engineering for us all to like write this to get our agents to do things. That would be terrible. So, we should try it. >> So, we're comparing no system prompt, unmannered, and plain writing. So, these are three different ones. All right, let's have a bake off.

2:17 >> Okay, so so far unmatered is is the loser for me, >> but the the bones are better here. Yeah, I can I would say let's call the C a winner. >> I agree. Yeah, I agree. >> I think I like B a little bit more. If you want to put a gun to my head, I would say okay, let's go with B. It is very interesting that the unmanned pros is so close to like a good writing promp or good writing prompt. What we thought was good. I think the unmannered one did a little bit better because it said 17 over 22 versus 14 over 22. That made it a little bit easier to know. And it also included one less fact about the cost, which is nice. And then like the finding GP versus semantic search, same accuracy, cheapest run is a lot better, I think, than GP versus semantic search over the same OCR output. Same accuracy, cheapest run. The bottleneck was never search. I mean, it's true, but it's like too wordy. It's not good.

3:06 >> Yeah, the slop is here. I don't know. I think B is the best. >> I think so too. >> I think what I'm learning so far is like there is no winner all the time necessarily. Like you could have good things from each. >> I think it's it's it's it's very interesting and surprising to me that there's not a clear winner. >> If I was to try to do this for a serious task, >> I might I might say like C I would say like C is a better bones to it. Was like Yeah. If I think about how I'm going to take a and then like edit it to be good, it's like I think Yeah.

3:38 >> Okay. Unmannered. So, I don't know, man. The unmannered it kind of is way stronger. I mean, we only looked at five examples, but I thought it would be like not >> I thought it would I thought it was just going to be a joke. I thought it was going to be terrible. Is it just that the prompt for writing doesn't really matter? Like, could we write a prompt that's like write like a good writer and is that just as good as all the other prompts?

4:01 >> No. This is inconclusive. This is not enough data to conclude anything. It just tells us that it's all over the place. >> And I I I learned that it's a lot closer than I thought. I thought for sure that the unmannered would just never be better. >> Yeah. I thought there was no way that's going I thought like, oh, like my own handcrafted prompt that's like has a lot of rules in it and that's like more carefully considered is going to be a lot better than than this slop. I'm not really sure that's the case. And I'm not sure like writing a prompt at all like in some cases writing any prompt is is like helpful. And the fact that you have >> this piece of slop and then they said this short version also tends to work which is just the word manner pros that's all it has. That does like kind of fit in with your hypothesis like maybe the prompt doesn't matter so much which is really interesting. Yeah. It's like really the context I think that was steering it the most. I was actually kind of skeptical about doing this because I was like, we're just going to come on here and it's going to be like a joke stream and talking about how this doesn't work and it's kind of obvious what's the value. But it surprised me.

Summary

The discussion centers around the effectiveness of different writing prompts for AI, particularly comparing "mannered prose," "unmannered prose," and plain writing. The participants explore how the structure and style of prompts influence the quality of AI-generated text, ultimately finding that the differences are less pronounced than expected.

- Claude's fables 5.1 shows improvement in writing quality, reducing stock phrases and jargon.
- Mannered prose often complicates writing, making it harder for readers to understand.
- The effectiveness of prompts may not be as significant as previously thought; context plays a crucial role.
- A comparison of unmannered and mannered prose reveals that unmannered prose can sometimes yield better results.
- Participants express skepticism about the need for complex prompts, suggesting that simpler instructions might suffice.
- The findings indicate that there is no definitive winner among the different prompt styles; effectiveness can vary based on context.
- The experiment challenges the assumption that well-crafted prompts always outperform simpler ones.

Questions Answered

What are the strengths and weaknesses of Claude's writing style?

Claude's writing has improved with fewer stock phrases and less jargon, but it can still be dense and difficult for readers. The use of mannered prose can obscure meaning and make reading harder.

How does the inclusion of sloppy writing in prompts affect AI output?

The expectation is that including sloppy writing in prompts would lead to more sloppy output, but initial findings suggest that it might not be the case.

What are the results of comparing unmannered, plain writing with other prompts?

The comparison reveals that there isn't a clear winner among the different writing styles, with each having its strengths and weaknesses.

What conclusions can be drawn about the effectiveness of writing prompts?

The analysis indicates that the effectiveness of writing prompts is not straightforward, and there may be valuable elements in each style.

Do prompts really matter in AI writing?

The findings suggest that the effectiveness of prompts may not be as critical as previously thought, with context potentially being more influential.

© transcribe · For agents Built with care and craft by Gokul Rajaram