zatkin

As someone who isn't a machine learning expert by any means, are these "System One" models effectively a "refreshed look" at prior art in the space, or is it truly novel new ways of using models? I'm deeply invested in trying to dissect more from this space, as I think inserting a small {classifier,decider,extractor,...} model at any point in the SDLC can drastically improve the state of affairs. More specifically, I'm thinking that since these models are so small, they're able to fit into Docker containers, which are effectively run ~everywhere for CI / CD, and cost basically whatever is the cost of storing a Docker image.

RuneScape was what led me to become a software engineer (through the private server communities) and it's refreshing to see someone describe this game in layman terms, all the way up through to having AI agents drive the player interactions. I worry, though, about Jagex's stance on all of this. Perhaps this is what pushed them towards the recent RuneScape 4 announcement?

How does it handle the scenario where the sender changes the domain they originate from? It may seem a bit unusual, like when a company rebrands (think of like Facebook originally sending from facebook.com, but then sending from meta.com later on).

Whose attention are you trying to capture by making this post?

Is it just me or does it feel like Anthropic and OpenAI (and probably others) are pumping harness engineering and the like to try and convince engineering organizations, or even entire companies, to "use the platform", and not meddle with software anymore?

It's working for me (based out of California).

This website is unusable without closing out advertisements upon advertisements...

I'm in the same boat as you, having only read non-fiction the past few years. It appears quite ironic that the entertainment occurring in the real life events is surpassing the entertainment I hear one usually derives from fiction.

I wonder what the dress code is.

The author mentions CADO-NFS right in the article: https://cado-nfs.gitlabpages.inria.fr/

This has got to be the worst project name I've seen all year.

I don't think I could ever go back to running an e-mail server again, given the continually evolving battle with convincing the major e-mail providers that your IP is legit (I'm looking at you AT&T!).

From a different angle, I moved from hosting my own e-mail server to having it all served in Fastmail, and I couldn't be happier as well!

I wonder if you put job classifications along a spreadsheet's leftmost column (e.g. search in file, search in directory), and data types along a spreadsheet's topmost row (e.g. JSON file, macOS-based filesystem), would you end up finding gaps in the intersection of the columns and rows make these new tools you've created less surprising? For example, to "search a file" on a "macOS-based filesystem", you would naturally gravitate towards `grep`. Or to search the contents of a JSON file, you would maybe use `grep`, or perhaps `jq` and a well crafted query. Well, as you expand out the job classifications and data types (or some other abstraction), you start to realize where we don't have tools today. And then Claude can go and create those, perhaps even proposing a faster alternative!

I've never installed the ChatGPT app on my Mac until today. I ran the app, signed in, and I see the same.

% ls ~/.cache/codex-runtimes/codex-primary-runtime/dependencies/native/libreoffice-headless

libreoffice manifest.json

I logged in just to share the entire comment for you:

>The big milestone is when computers can read and understand information like humans do. There is a lot of work going on in this field - Google, Microsoft, Facebook, academia,... Right now computers don't know how to represent knowledge so they can't read a text book and pass a test.

>Another whole area is vaccines. We need a vaccine for HIV, Malaria and TB and I hope we have them in the next 10-15 years.

Maybe I'm being super reductive here, but operating small models at the core of your business kind of moves the needle from making external API calls (against frontier models) to running internal API calls (against your locally-run models). It seems like if we want local models to take off, it will need to become easier to run local models for cheap. I'm thinking like reducing the barrier of entry for running "local models" in the cloud providers like DigitalOcean, AWS, etc.

Doesn't this skill only work up until Anthropic tweaks weights or the system prompt(s) to not allow it?

I personally own a Supernote, but I'm not a heavy user of it. For the sake of my own curiosity, what benefits will you get out of having reverse engineered the Supernote note file format? It would be super rad to be able to move my notes between other devices, which is one big plus that comes to my mind.

There's some real research-backed evidence that this approach elevates output quality, and it was even discussed a few months back: https://news.ycombinator.com/item?id=47169627

I'm the horrible kind of person to go off and use an AI-assisted cheat that can be pasted into the console. Very fun game, nonetheless!

I find it hard to buy this generalization about _all_ "AI"/agent software when the author has only provided _one_ example. A couple alternatives could've been explored, like Cursor Cloud Agents or even just running Claude Code with remote control on another machine. I'd also personally never spend a dime on any product unless I've read _some_ positive sentiment from online commentary, like Hacker News. To my point, Ona appears to have a pretty low amount of feedback that makes it difficult to justify any $20 bills: https://hn.algolia.com/?q=ona.com

I tend to ask the LLM for a single HTML page explanation, with a pedagogical approach. Something about dropping the word pedagogical leads to a more structured outcome, but I haven't quite figured it out why yet.

>This project is a fully working native Windows x64 port of Microsoft Word for Windows 1.1a, whose historical codename was Opus.

Not to be confused with the Anthropic LLM model Opus - what was once an old codename is as good as new once again!

The emdashes from your LLM-generated responses are leaking. I would encourage you to consider writing comments here as your own, which better reflects you as a person.