tech

This link was posted by [kelseyfrog](https://news.ycombinator.com/user?id=kelseyfrog) 1 hour ago on [HN](https://news.ycombinator.com/item?id=49783495). It received 107 points and 20 comments.

[Top comment](https://news.ycombinator.com/item?id=49783677) by [fredrikholm](https://news.ycombinator.com/user?id=fredrikholm):

> I found Grim Fandango in a pile of discounted games as a tween in the 2000s. There was a cool skeleton in a suit on the front cover, so I bought it based on that.
> 
> Even though I didn't understand most references and sarcasm, few games have grabbed my attention as Grim Fandango did. The art work, the music, the writing. The whole game oozes of style that I've never seen replicated.
> 
> Even some 20+ years later I can almost recite most of Act I from heart. If you haven't played this game and like adventure games, this is one of the best.
> 
> Thank you for linking OP, this made my morning.

Grim Fandango Puzzle Document (1996) [pdf]

This link was posted by [blazarquasar](https://news.ycombinator.com/user?id=blazarquasar) 2 hours ago on [HN](https://news.ycombinator.com/item?id=49780797). It received 121 points and 47 comments.

[Top comment](https://news.ycombinator.com/item?id=49781645) by [prng2021](https://news.ycombinator.com/user?id=prng2021):

> Can someone clarify the use case for this? What's the benefit over this:
> [https://openai.com/index/introducing-the-agents-api/](https://openai.com/index/introducing-the-agents-api/)

Google's Open Agentic Orchestrator

This link was posted by [giuliomagnifico](https://news.ycombinator.com/user?id=giuliomagnifico) 2 hours ago on [HN](https://news.ycombinator.com/item?id=49778029). It received 139 points and 100 comments.

[Top comment](https://news.ycombinator.com/item?id=49778660) by [fooker](https://news.ycombinator.com/user?id=fooker):

> What's the main blocker (other than the current inflated cost) for using HBM instead of DRAM as the primary memory for consumer electronics?

Samsung is expected to more than double output of its HBM4 and HBM4E DRAM

This link was posted by [skepticalgenius](https://news.ycombinator.com/user?id=skepticalgenius) 2 hours ago on [HN](https://news.ycombinator.com/item?id=49776699). It received 173 points and 61 comments.

[Top comment](https://news.ycombinator.com/item?id=49777594) by [wren6991](https://news.ycombinator.com/user?id=wren6991):

> Speaking to the "uncensored model" angle: there's little reason to distribute abliterated weights anyway. Instead of orthogonalising the weights that write back to the residual stream, you can just orthogonalise the activations themselves. It's equivalent.
> 
> Orthogonalising activations at runtime is computationally cheap. Just distribute the refusal vectors (few thousand floats per layer), then run against the stock weights. Antirez's DS4 already supports this: [https://github.com/antirez/ds4/blob/8db1d1d155cb0400a86a86b9...](https://github.com/antirez/ds4/blob/8db1d1d155cb0400a86a86b9c62d0defb3a6148b/dir-steering/README.md)
> 
> …

Pirate Face Rescues LLM Models from Deletion

This link was posted by [rakel_rakel](https://news.ycombinator.com/user?id=rakel_rakel) 2 hours ago on [HN](https://news.ycombinator.com/item?id=49774329). It received 107 points and 52 comments.

[Top comment](https://news.ycombinator.com/item?id=49775094) by [grumbel](https://news.ycombinator.com/user?id=grumbel):

> I really don't get those takes. AI is the best thing that ever happened to the Free Software world, it is basically turning any software into Free Software. You can just throw file formats, protocols or even plain binaries at the AI and it'll reverse engineer everything in a pinch. Users can finally modify software themselves, which was always the goal of the Free Software world, but very rarely happened in actuality, since it was just so damn complicated. AI lowered the barrier of entry tremendously, not just in terms of required knowledge, but especially time. Same with Creative Commons, sharing art and stuff, was a nice gesture, but rarely useful, since the level of work to modify a work to fit your project was pretty close to just doing it from scratch anyway. With AI everybody can toy around with image generators and get what they want.
> 
> …

AI and the Destruction of the Creative Commons

This link was posted by [RohanAdwankar](https://news.ycombinator.com/user?id=RohanAdwankar) 1 hour ago on [HN](https://news.ycombinator.com/item?id=49771110). It received 113 points and 45 comments.

[Top comment](https://news.ycombinator.com/item?id=49771331) by [AceJohnny2](https://news.ycombinator.com/user?id=AceJohnny2):

> I haven't bothered to test the API, but you've effectively allowed a fully-open upload API? Who's paying the storage costs, and how do you prevent abuse?
> 
> (Obviously I'm taking this more seriously than it's probably meant to)

Exfiltrate Your Weights

This link was posted by [theanonymousone](https://news.ycombinator.com/user?id=theanonymousone) 2 hours ago on [HN](https://news.ycombinator.com/item?id=49770293). It received 109 points and 58 comments.

[Top comment](https://news.ycombinator.com/item?id=49770965) by [kens](https://news.ycombinator.com/user?id=kens):

> Author here. I don't know why this popped up 13 years later, but hi :-)

How Hacker News ranking works: scoring, controversy, and penalties (2013)

This link was posted by [nandakishor_ml](https://news.ycombinator.com/user?id=nandakishor_ml) 3 hours ago on [HN](https://news.ycombinator.com/item?id=49765348). It received 227 points and 35 comments.

[Top comment](https://news.ycombinator.com/item?id=49765997) by [Oras](https://news.ycombinator.com/user?id=Oras):

> I played around with Jev last night and did it for classification tasks that I used Gemini 2.5 flash lite with.
> 
> It’s a bit faster and bit cheaper, but this is compared to LLM. The consistency was nice to see, BUT, as someone who trained NLP models prior to LLMs, it’s just BERT with more data. I can see why people would want ready made one shot classifier, and I can see the value of sending multiple classifier in one call, but I wouldn’t call it breakthrough. And I believe many labs will replicate it in no time and might have it as part of their harness.
> 
> I see it as a wake up call for the tech community to go back to basics for most tasks instead of relying solely on generic LLMs.

Laya the open source version of Jev

This link was posted by [ereiamjh](https://news.ycombinator.com/user?id=ereiamjh) 3 hours ago on [HN](https://news.ycombinator.com/item?id=49764791). It received 301 points and 190 comments.

[Top comment](https://news.ycombinator.com/item?id=49765646) by [ajjenkins](https://news.ycombinator.com/user?id=ajjenkins):

> A lot of people in the comments say that the supposedly better examples in the article are still “horrible” because they can tell they’re AI and point to things that the AI did “wrong” and a human artist would do better.
> 
> That may be true if you’re working with a top of the line artist, but in my experience the average freelance designer online is significantly worse than AI. I’ve worked with dozens of budget friendly designers on Fiverr and consistently get way worse results than I’ve gotten from AI models. Will a great human designer produce better output than an AI? Maybe, but a small local event or business probably won’t have the budget to hire them or know where to find them. It’s cheaper, faster and easier to use AI to generate something that’s good enough for their purposes.
> 
> …

AI-generated posters don’t have to be horrible

This link was posted by [emigre](https://news.ycombinator.com/user?id=emigre) 1 hour ago on [HN](https://news.ycombinator.com/item?id=49763697). It received 114 points and 30 comments.

[Top comment](https://news.ycombinator.com/item?id=49764026) by [itvision](https://news.ycombinator.com/user?id=itvision):

> What a way to oversell the discovery.
> 
> The reality is:
> 
>     * Different brain regions have different functions: ancient knowledge.
>     * Different regions contain distinct cell types and molecular programs: also well established.
>     * Anterior and posterior brain structures may trace back to fundamentally separate progenitor lineages that were independently specified very early in evolution and development: that's the interesting new result.
> 
> The headline should have been: "Human forebrain and hindbrain arise from distinct embryonic progenitor lineages" - not that we have "two organs" in the brain.

Human brain is two separate organs, Stanford Medicine-led research finds

This link was posted by [z-mach9](https://news.ycombinator.com/user?id=z-mach9) 1 hour ago on [HN](https://news.ycombinator.com/item?id=49763296). It received 119 points and 41 comments.

[Top comment](https://news.ycombinator.com/item?id=49763509) by [mimorigasaka](https://news.ycombinator.com/user?id=mimorigasaka):

> The people behind this are a real (if silly) student organization at UChicago and Northwestern campaigning for the legalization of onion futures trading. Sometimes they will aggressively hand out free onions to passerby on campus. It's quite entertaining.
> 
> They have a website also: [https://onionfutures.org](https://onionfutures.org)

San Francisco Onion Futures Company

This link was posted by [datadrivenangel](https://news.ycombinator.com/user?id=datadrivenangel) 1 hour ago on [HN](https://news.ycombinator.com/item?id=49760187). It received 272 points and 110 comments.

[Top comment](https://news.ycombinator.com/item?id=49760745) by [rukuu001](https://news.ycombinator.com/user?id=rukuu001):

> Once I asked, puckishly, Claude Code to “follow the instructions in this directory” when there was only an AGENTS.md there.
> 
> In the manner of someone finding a dead mouse and holding it up for examination CC said it could find no instructions but perhaps it should check this AGENTS.md file.

Claude Code now reads AGENTS.md if there is no Claude.md

This link was posted by [theanonymousone](https://news.ycombinator.com/user?id=theanonymousone) 2 hours ago on [HN](https://news.ycombinator.com/item?id=49758736). It received 228 points and 105 comments.

[Top comment](https://news.ycombinator.com/item?id=49759227) by [wps](https://news.ycombinator.com/user?id=wps):

> The amount of roadblocks Google is putting up for GrapheneOS is just ridiculous. None of their decisions make any sense, from the delayed source patches upstream, to the embargos, attestation issues, etc. Google simply regrets android being open source.

Android 17 is the first since 3.x to add new APIs without releasing to the AOSP

This link was posted by [wek](https://news.ycombinator.com/user?id=wek) 2 hours ago on [HN](https://news.ycombinator.com/item?id=49753878). It received 119 points and 20 comments.

[Top comment](https://news.ycombinator.com/item?id=49755993) by [lieret](https://news.ycombinator.com/user?id=lieret):

> Cool study, we definitely need more principled studies on the role of harnesses. I'd also say that there aren't too many benchmarks where the more complicated harnesses consistently outperform extremely simple agents. But I'm also biased, because I wrote [https://github.com/swe-agent/mini-swe-agent/](https://github.com/swe-agent/mini-swe-agent/) , which is probably the most minimal agent out there (it started as just 100 lines, all included), and it's used in a lot of benchmarks like DeepSWE, terminalbench, programbench (seems like it's still top of the ranking for TB3, but wasn't evaluated with the best models on TB4).

An Empirical Study of Harness Design for Coding Agents

This link was posted by [LiamPowell](https://news.ycombinator.com/user?id=LiamPowell) 1 hour ago on [HN](https://news.ycombinator.com/item?id=49753179). It received 151 points and 77 comments.

[Top comment](https://news.ycombinator.com/item?id=49753379) by [pu_pe](https://news.ycombinator.com/user?id=pu_pe):

> The original discussion about the project ([https://news.ycombinator.com/item?id=49746163](https://news.ycombinator.com/item?id=49746163)) is very weird. Lots of call-outs about how the author is some sort of celebrity and random accounts vouching for him, with little discussion on the substance.
> 
> Not even the demo on that release works well.

Bend 2 and the Vibe-Coding Trap

This link was posted by [ilreb](https://news.ycombinator.com/user?id=ilreb) 1 hour ago on [HN](https://news.ycombinator.com/item?id=49752041). It received 118 points and 57 comments.

[Top comment](https://news.ycombinator.com/item?id=49752729) by [druskacik](https://news.ycombinator.com/user?id=druskacik):

> I'm really interested in technical details behind Jev (not this), how it can work so fast and so cheap. It's probably large (must be since the performance is so good) but somehow still fast, so it must include some really non-trivial stuff. The price suggests it may be runnable locally, but who knows.
> 
> If it was possible to re-create it as an open-weight, it would be exciting!

OpenJev

This link was posted by [gkfasdfasdf](https://news.ycombinator.com/user?id=gkfasdfasdf) 4 hours ago on [HN](https://news.ycombinator.com/item?id=49750152). It received 121 points and 34 comments.

[Top comment](https://news.ycombinator.com/item?id=49750698) by [vocx2tx](https://news.ycombinator.com/user?id=vocx2tx):

> For some context on why this release is notable: Jemalloc Postmortem [0]
> 
> [0] [https://news.ycombinator.com/item?id=44264958](https://news.ycombinator.com/item?id=44264958)

Jemalloc 5.4.0

This link was posted by [Handy-Man](https://news.ycombinator.com/user?id=Handy-Man) 2 hours ago on [HN](https://news.ycombinator.com/item?id=49749656). It received 129 points and 36 comments.

[Top comment](https://news.ycombinator.com/item?id=49750117) by [oefrha](https://news.ycombinator.com/user?id=oefrha):

> Unsandboxed ImageMagick is known for being a security nightmare even back when PHP ruled the world (not saying sandboxing is a panacea either, it just requires a different and potentially harder exploit to develop a full chain). Difference is it's easier than ever to turn vulnerabilities into full compromises. At some point we'll have to replace all parsers with something at least as safe as [https://github.com/google/wuffs](https://github.com/google/wuffs) right? Otherwise ImageMagick and co. will just keep giving.

Hacking OpenAI

This link was posted by [vertigoruntime](https://news.ycombinator.com/user?id=vertigoruntime) 1 hour ago on [HN](https://news.ycombinator.com/item?id=49745940). It received 146 points and 122 comments.

[Top comment](https://news.ycombinator.com/item?id=49746363) by [piker](https://news.ycombinator.com/user?id=piker):

> Second paragraph:
> 
> > API customers including Harvey and Legora will be able to build on Astra for Law, bringing this intelligence into their own products and workflows.
> 
> In other words: "no, no, we're not eating our children to prep for the IPO. Don't worry."

Astra for Law

This link was posted by [bookofjoe](https://news.ycombinator.com/user?id=bookofjoe) 3 hours ago on [HN](https://news.ycombinator.com/item?id=49743097). It received 209 points and 74 comments.

[Top comment](https://news.ycombinator.com/item?id=49744288) by [asciimoo](https://news.ycombinator.com/user?id=asciimoo):

> Ohi, author here! Thanks for posting Hister. Feel free to A.M.A.
> My first free software search project was Searx, a privacy respecting metasearch engine, but because of the limitations of the metasearch concept, I've decided to take a different approach.
> 
> Hister builds a personal search index from pages you visit, bookmarks, browser history, local files, and crawled websites. It stores extracted content with offline result previews, so information remains searchable even when the original page changes or disappears. It supports full text and semantic search, can run entirely on your own machine, and includes a web interface, command line tools, and an MCP endpoint for assistant integrations.
> 
> Website: [https://hister.org/](https://hister.org/)
> 
> Tiny read-only demo: [https://demo.hister.org/](https://demo.hister.org/)
> 
> …

Hister: A private search engine for the pages you visit and the files you keep

This link was posted by [my123](https://news.ycombinator.com/user?id=my123) 1 day ago on [HN](https://news.ycombinator.com/item?id=49715813). It received 126 points and 45 comments.

[Top comment](https://news.ycombinator.com/item?id=49740724) by [alightsoul](https://news.ycombinator.com/user?id=alightsoul):

> Fun fact, Fujitsu used to run their own fabs in Japan until they were sold to UMC and now they're another UMC fab although they are stuck at like the 20nm node. So this is likely made at JASM.