It's interesting that you chose to measure how well each LLM did in talking to other NPCs, and having each NPC also use an LLM to react to each input. Why not have the LLM fight NPCs and loot items (both a significant part of the MUD experience, neither requiring LLMs on the server side), then measure character progression in experience and stats?
HN user
aphexairlines
daniel@danielsilva.org
http://www.amazon.com/gp/aw
Also standalone: https://www.meta.ai/ and https://about.fb.com/news/2025/04/introducing-meta-ai-app-ne...
If your requirements.in referenced numpy before this release, then doesn't your requirements.txt already reference a specific 1.x version?
Not a Thinkpad and not Intel, but that already seems attainable with the asus G14 (AMD 8945HS + NVIDIA 4070) or the upcoming asus zenbook S16 (AMD HX 370).
I've never seen a large company with more L7+ engineers than L7+ managers. Usually it's a ratio of drastically more managers hired/promoted into those levels than engineers, and some orgs don't have any engineers in those levels at all.
Take a look at https://github.com/xonsh/xonsh/issues before deciding to abandon the devil you know.
I prefer sticking with bash where necessary (where a script is the only thing that will reasonably work), and elsewhere using a programming language with testing, type checking, modularity, and compilation into something with zero or minimal runtime dependencies.
This is a JavaScript keyword and there's nothing TypeScript-specific in the blog post.
Is this similar to people moving away from giving their code as open source when companies take it, repackage, and sell it while the authors get nothing?
Looks like fake reviews here too: https://news.ycombinator.com/item?id=34968995
They don't have to hire people who already know Zig -- just people capable of learning it.
You don't need to run jlink yourself to avoid shipping the entire JDK. Here's a java19 runtime docker image at 62MB: https://hub.docker.com/_/eclipse-temurin/tags?page=1&name=19...
Not customizing the runtime per build or per app also potentially helps reuse of docker image layers across your container registries and clusters.
You still need postcss or something like it to use css modules. The alternative of CSS-in-JS requires a build step to extract either a CSS file or inline critical styles for server rendering. The browser-native alternative of CSS Module Scripts doesn't work in Safari.
You still need babel, tsc, or swc to transform typescript to JS.
And then there are imports of files that aren't JS (images, fonts, stylesheets, etc).
There's still no good way of publishing an npm package with multiple components, stylesheets per component, and transitive consumers using only the stylesheets for the components they reference while rendering fully server side. CSS is still not scalable.
Very good to have a smaller runtime. Some of the stated benefits of the new generator are silly and the comparison with ts-proto doesn't point out the actual improvement (generating d.ts files means your CI doesn't spend time type-checking and compiling ts sources), but nice to see this new alternative.
Comments explain why the code is there. That's different from translating code into English.
Or just walk down Peterborough street: https://elpelon.com
There's a super interesting article from the just-js author explaining how he got a JS implementation into that top 20: https://just.billywhizz.io/blog/on-javascript-performance-01...
- High performance
- Scripting language
Pick one.
I think that's ocaml. Higher level, with GC, nice type system, no traits but signatures and higher order modules might be enough for you, and a compiler that produces fast native binaries.
Yes. Even this week, we have the news that his supreme court nominees are voting to strike down abortion protections, and I have female friends and family in states with trigger laws.
create-react-app starts with 1506 deps:
$ npx create-react-app my-app
$ find my-app/node_modules -name package.json | wc -l
1506in my life
Lucky you?
Whoever sold the vulnerable software should be liable for damages.
It's also easy to provide a default factory: https://doc.rust-lang.org/std/default/trait.Default.html
Reconsider because you'll do yourself a disservice by expecting the world to work that way. You don't want to be surprised and upset when your employer rewards someone for introducing more bureaucracy instead of some technically challenging work you delivered.
If you truly believe this, then please reconsider. The tech industry doesn't reward people for technically challenging work. Consider how many large companies were/are built on what you'd consider to be terrible technology choices.
Don't expect an "ultimately necessary transition" because it'll likely go in the opposite direction.
Health service in the US comes with liability:
https://www.ncsl.org/research/financial-services-and-commerc...
The HHS regulates public health at the federal level. It has around 80,000 employees.
https://www.hhs.gov/regulations/index.html
https://en.wikipedia.org/wiki/United_States_Department_of_He...
the invisible hand
The FDA has 18,000 employees. The USDA has around 100,000.
Given the same salaries in both countries, it's probably more expensive to hire that employee in the US given that the US firm is expected to pay for fairly expensive health insurance.
UK firms pay for private health insurance too, but that's supplemental insurance on top of NHS coverage.
Java's not helpful for FP, but that's an odd example to bring up in a discussion about Go, which is also statement-oriented.