Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

The new skill in AI is not prompting, it's context engineering

philschmid.de

Post reply on HN
Print view View original

521–530 of 550 posts

Re: The new skill in AI is not prompting, it's context engineering

#521

Post by groestl »

shakna wrote: post #315
groestl wrote: post #282

Earlier quoted context omitted.

Well, to get the right thing into the context in a performant way when you dealing with a huge dataset is definitely engineering.

Engineering tends to mean "the application of scientific and mathematical principles to practical ends". I'm not sure there's much scientific or mathematical about guessing how a non-deterministic system will behave.

Funny how you use a scientific term to discredit applied statistics. I've built useful non-deterministic systems many times and they had nothing to do with AI. Also, particle physics would like to have a word with you.

Re: The new skill in AI is not prompting, it's context engineering

#522

Post by groestl »

grugagag wrote: post #411
SonOfLilit wrote: post #359

Earlier quoted context omitted.

The moment you start building evaluation pipelines and running experiments to validate your ideas it stops being guessing

It’s validated and filtered but isn’t it still guessing at the core? Should we call it validated guessing?

I think a validated guess is exactly what a prediction is.

Re: The new skill in AI is not prompting, it's context engineering

#523

Post by groestl »

ModernMech wrote: post #445
SonOfLilit wrote: post #359

Earlier quoted context omitted.

The moment you start building evaluation pipelines and running experiments to validate your ideas it stops being guessing

You've heard of science versus pseudo-science? Well.. Engineering: "Will the bridge hold? Yes, here's the analysis, backed by solid science." Pseudo-engineering: "Will the bridge hold? Probably. I'm not really sure; although I have validated the output of my Rube Goldberg machine, which is supposedly an expert in bridges, and it indicates the bridge will be fine. So we'll go with that." "prompt engineer" or "context…

Engineering: "Will the bridge hold? Yes, with a confidence of 99.95%"

Re: The new skill in AI is not prompting, it's context engineering

#524

Post by Leo-thorne »

I learned this the hard way. Even a great prompt won't work if the context window is off. If key information is missing or important history is buried too deep, the model will still fail. Now I always explain the problem clearly and set the scene before letting the AI take over.

Re: The new skill in AI is not prompting, it's context engineering

#525

Post by Daub »

skydhash wrote: post #519
Daub wrote: post #510

Earlier quoted context omitted.

I concur that there is, on some matters, a general agreement in art books. However, certainly it does not help that there is so much inconsistency of terminology. For example: the way that hue and color are so frequently used interchangeably, likewise lightness, brightness, tone and value. What bothers me more is that so much truly important material is not being addressed as explicitly as it should be. For example:…

> For example, a common device that painters employ is to configure the neighboring regional contrast of a form can be light against dark on one edge and dark against light on the opposing edge. I'm not fully sure of what you means. If we take the following example, are you talking about the neck and the collar of the girl? https://i.pinimg.com/originals/ea/70/0b/ea700b6a0b366c13187e... https://fr.pinterest.com/pin/4…

> I'm not fully sure of what you means. If we take the following example, are you talking about the neck and the collar of the girl?

Yes... that's exactly it. It is also described in our teaching material here, (half way down the page):

https://rmit.instructure.com/courses/87565/pages/structural-...

Rembrandt was an avid user of this technique. In his portraits, one little trick he almost always used was to ensure that there was no edge contrast whatsoever in at least one region, usually located near the bottom of the figure. This served to blend the figure into the background and avoid the flat effect that would have happened had he not used it. In class I call this 'edge loss'. An equivalent in drawing is the notion of 'open lines' whereby silhouette lines are deliberately left open at select points.

> I think the name of the concept is "edge control" (not really original). You can find some explanation here.

I am aware of the term 'edge control' though I have not heard it used in this context. I feel that the term is too general to describe what is happening in the (so-called) tone wrap.

To extend the principle, wrap is an important concept in spatial rendering (painting, photography, filmmaking etc) and is a cousin of overlap. Simply... both serve to enhance form.

> To keep it short, there's no line in reality.

True that. I learned a lot about lines from reading about non-photorealistic rendering in 3D. There are some great papers on this subject (below) though I feel there is still work to be done.

Cole, Forrester, et al. "How well do line drawings depict shape?." ACM SIGGRAPH 2009 papers. 2009. 1-9.

Cole, Forrester, et al. "Where do people draw lines?." ACM SIGGRAPH 2008 papers. 2008. 1-11.

I made a stab at summarizing their wisdoms here:

https://rmit.instructure.com/courses/87565/pages/drawing-lin...

> There's a book called "Color and Light" by James Gurney that goes in depth about all of these. There's a lot of parameters that goes inside a brush stroke in a specific area of a painting.

Looking at it now. Any writer who references the Hudson River School is a friend of mine.

Re: The new skill in AI is not prompting, it's context engineering

#526

Post by hackable_sand »

zacharyvoase wrote: post #99

I love how we have such a poor model of how LLMs work (or more aptly don't work) that we are developing an entire alchemical practice around them. Definitely seems healthy for the industry and the species.

This is offensive to alchemy.

Re: The new skill in AI is not prompting, it's context engineering

#527

Post by hackable_sand »

jm4 wrote: post #415
TrackerFF wrote: post #395

It is probably 6-7 months ago I used ChatGPT for "vibe coding", and my main complaint was that the model eventually started moving away too far from its intended goal, as and it eventually go lost and stuck in some loop. In which case I had to fire up a new model, and feed all the context I had, and continue. A couple of days ago I fired up o4-mini-high, and I was blown away how long it can remember things, how much…

I went down that rabbit hole with Cursor and it's pretty good. Then I tried tools like Cline with Sonnet 4 and Claude Code. The Anthropic models have huge context and it shows. I'm no expert, but it feels like you reach a point where the model is good enough and then the gains are coming from the context size. When I'm doing something complex, I'm filling up the 200k context window and getting solutions that I just c…

Recently cancelled my Cursor subscription because it's pretty bad.

Re: The new skill in AI is not prompting, it's context engineering

#529

Post by fwn »

simonw wrote: post #403
fwn wrote: post #388

Earlier quoted context omitted.

That’s pretty typical, though not especially reliable. (Allthough in my experience, Gemini currently performs slightly better than ChatGPT for my case.) In one repetitive workflow, for example, I process long email threads, large Markdown tables (which is a format from hell), stakeholder maps, and broader project context, such as roles, mailing lists, and related metadata. I feed all of that into the LLM, which deter…

Try reformatting the data from the markdown table into a JSON or YAML list of objects. You may find that repeating the keys for every value gives you more reliable results.

Thanks for the suggestion! I’ll start benchmarking my current md table setup against one using YAML. It's apparently slightly less verbose than JSON.
Post reply on HN