frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Defining and evaluating political bias in LLMs

https://openai.com/index/defining-and-evaluating-political-bias-in-llms/
1•stefanpie•1m ago•0 comments

Meta's new display-less smart glasses are quite good, but the vibes are off

https://www.wired.com/review/ray-ban-meta-gen-2-glasses/
1•CharlesW•2m ago•0 comments

Show HN: AI agents running on 2011 Raspberry Pi with pure PHP – no GPU

https://github.com/paolomulas/datapizza-ai-php
1•paolomulas•2m ago•0 comments

The declining population will make it even harder to care for elders

https://text.npr.org/nx-s1-5535648
2•mooreds•3m ago•0 comments

Show HN: Open in Cursor" Finder Extension for macOS

https://github.com/inem/OpenInCursor
1•inem•3m ago•0 comments

Turing Patterns in Photoshop (2015) [pdf]

https://archive.bridgesmathart.org/2015/bridges2015-459.pdf
2•o4c•5m ago•0 comments

OpenAI co-founder: AI agents are still 10 years away

https://thenewstack.io/openai-co-founder-ai-agents-are-still-10-years-away/
1•MilnerRoute•6m ago•0 comments

The Stupidly Simple Medium Hack That's Generating Page 1 Rankings in 2.5 Months

https://larslofgren.com/medium-affiliate-scam/
1•caminanteblanco•8m ago•0 comments

When execution gets easy, taste gets harder

https://www.antonsten.com/articles/when-execution-gets-easy-taste-gets-harder/
2•janpio•9m ago•0 comments

RP2350 UART Driver in Assembler

https://github.com/mytechnotalent/RP2350_UART_Driver
1•mytechnotalent•11m ago•0 comments

Transformers Explained: The Discovery That Changed AI Forever [video]

https://www.youtube.com/watch?v=JZLZQVmfGn8
1•gmays•11m ago•0 comments

Wanggongchang Explosion

https://en.wikipedia.org/wiki/Wanggongchang_Explosion
1•thunderbong•12m ago•0 comments

The $38T Question: An Interview with Stanford Professor Hanno Lustig

https://stanfordreview.org/the-38-trillion-question-an-interview-with-stanford-professor-hanno-lu...
1•cjbarber•13m ago•0 comments

Show HN: AIs, 1 religion: what my experiment revealed about AI bias

1•Anh_Nguyen_vn•13m ago•0 comments

Ask HN: Are world models going to be the next big thing?

1•pranoy•20m ago•0 comments

Show HN: Get AI job tools and Premium Perks at RemotelyGood.us starting at $5

https://remotelygood.us
1•Theresa_i_a•20m ago•0 comments

OpenGeom Fork for OpenCascade

https://quaoar.su/blog/page/opengeom-fork-for-opencascade
1•alangibson•20m ago•0 comments

Palantir Thinks College Might Be a Waste. So It's Hiring High-School Grads

https://www.wsj.com/business/palantir-thinks-college-might-be-a-waste-so-its-hiring-high-school-g...
13•adwmayer•20m ago•14 comments

Numb at Burning Man

https://samkriss.substack.com/p/numb-at-burning-man
2•lores•22m ago•0 comments

New way to organically market your product

https://rankmochi.com
1•gintokinx•24m ago•0 comments

How to get the GOT address from a PLT stub using GDB

https://rafaelbeirigo.github.io/cybersec-dojo/research/2025/11/01/how-to-get-the-got-address-from...
2•rafaelbeirigo•24m ago•0 comments

How to make more money with your employee stock options at no extra risk

http://www.weidai.com/stock-options.txt
1•leoh•27m ago•0 comments

Readweb: Extract token-efficient Markdown from websites

https://github.com/promptware/readweb
1•klntsky•29m ago•0 comments

Product Designer's workflow for prototyping with Cursor

https://hvpandya.com/vibe-coding
1•hemezh•30m ago•0 comments

How a CVD Diamond Is Made

https://www.youtube.com/watch?v=uE_Qnsh1_2A
1•akshatjiwan•32m ago•0 comments

The Toad Report #2

https://willmcgugan.github.io/toad-report-2/
1•willm•33m ago•0 comments

ZkML Breakthrough: 13B Models Verified in 15 Minutes

https://lightcapai.medium.com/decentralized-ai-systems-cryptographic-infrastructures-verifiable-c...
1•HenryAI•33m ago•1 comments

The Bear Manifesto: Longevity, planning for the future, legal structure

https://herman.bearblog.dev/manifesto/
2•gregwolanski•35m ago•0 comments

Meltwater Pulse 1B

https://en.wikipedia.org/wiki/Meltwater_pulse_1B
1•keepamovin•35m ago•0 comments

Hierarchy Elevates Social Reasoning

https://positron.solutions/articles/hierarchy-elevates-social-reasoning
1•positron26•36m ago•0 comments
Open in hackernews

Context engineering

https://chrisloy.dev/post/2025/08/03/context-engineering
49•chrisloy•7h ago

Comments

elteto•3h ago
Are there any open source examples of good context engineering or agent systems?
calebkaiser•39m ago
Any of the "design patterns" listed in the article will have a ton of popular open source implementations. For structured generation, I think outlines is a particularly cool library, especially if you want to poke around at how constrained decoding works under the hood: https://github.com/dottxt-ai/outlines
voidhorse•3h ago
There is nothing precise about crafting prompts and context—it's just that, a craft. Even if you do the right thing and check some fuzzy boundary conditions using autoscorers, the model can still change out from beneath you at any point and totally alter the behavior of your system. There is no formal language here. After all, mathematics exists because natural language is notoriously imprecise.

The article has some good practical tips and it's not on the author but man I really wish we'd stop abusing the term "engineering" in a desperate attempt to stroke our own egos and or convince people to give us money. It's pathetic. Coming up with good inputs to LLMs is more art than science and it's a craft. Call a spade a spade.

qrios•3h ago
I agree with you one hundred percent.

But: Interestingly, the behavior of LLMs in different contexts is also the subject of scientific research.

satisfice•3h ago
My thoughts exactly. The author is saying we should think strategically about the use of context. Sure. Yes. But for that to qualify as engineering we need solid theory about how context works.

We don’t have that, yet. For instance experiments show that not all parts of the context window are equally well attended. Imagine trying to engineer a bridge when no one really knows how strong steel is.

skeeter2020•2h ago
or how wide the river is year round
chrisweekly•2h ago
"Context crafting", ok, sure. I think a lot of expert researchers (like simonw) would agree.
grigio•3h ago
I'd like a RSS feed of this blog..
vladsanchez•1h ago
It's available, https://buttondown.com/chrisloy/rss but it's not in sync with the blog, just a single 2024 entry found. :shrug:
aeve890•2h ago
Are we still calling this things engineering?
skeeter2020•2h ago
"professionally trained & legally responsible for the results" is definitely not the same thing as what we used to just call "good at googling".
aeve890•1h ago
I'd say this shit is even worse that "good at googling". Literal incantation for stochastic machines is like just two notches above checking the horoscope.
calebkaiser•47m ago
Based on the comments, I expected this to be slop listing a bunch of random prompt snippets from the author's personal collection.

I'm honestly a bit confused at the negativity here. The article is incredibly benign and reasonable. Maybe a bit surface level and not incredibly in depth, but at a glance, it gives fair and generally accurate summaries of the actual mechanisms behind inference. The examples it gives for "context engineering patterns" are actual systems that you'd need to implement (RAG, structured output, tool calling, etc.), not just a random prompt, and they're all subject to pretty thorough investigation from the research community.

The article even echoes your sentiments about "prompt engineering," down to the use of the word "incantation". From the piece:

> This was the birth of so-called "prompt engineering", though in practice there was often far less "engineering" than trial-and-error guesswork. This could often feel closer to uttering mystical incantations and hoping for magic to happen, rather than the deliberate construction and rigorous application of systems thinking that epitomises true engineering.

sgt101•2h ago
Why would I believe that any of this works? This is just some blokes idea of what people should do.

There is no evidence offered. No attempt to measure the benefits.

calebkaiser•33m ago
Most of the inference techniques (what the author calls context engineering design patterns) listed here originally came from the research community, and there are tons of benchmarks measuring their effectiveness, as well as a great deal of research behind what is happening mechanistically with each.

As the author points out, many of the patterns are fundamentally about in-context learning, and this in particular has been subject to a ton of research from the mechanistic interpretability crew. If you're curious, I think this line of research is fascinating: https://transformer-circuits.pub/2022/in-context-learning-an...