blueblazin

Google at least has been investing very heavily into cloud security. They acquired Mandiant ($5.4 billion) and Wiz ($32 billion). Many people felt those were very expensive acquisitions at the time, but maybe it seems like a wise move now.

This was always one of the main reasons for accelerating AI. I really hope there is soon a moment for AI drug development similar to the recent AI math advances.

I had the code in a separate private repo because it had some personal data I hadn't cleaned up. When I shared this repo I forgot it didn't have the code! Sorry about that. A case of Hanlon's Razor and nothing more. I have uploaded the full codebase to the shared repo now.

No plans on monetizing it! Just made it for fun and have uploaded the full codebase now. I was using that repo for hosting the macOS download and had forgotten to host the codebase on it too! Sorry about that.

My apologies! I worked on this project a while back and had kept the code repo private because it contained some personal data I hadn't cleaned up. I had this repo just to host the macOS binary. When I shared it I forgot it had no code!

I have now uploaded the entire codebase to the repo! Neither sneaky nor deceiving, just plain old stupidity, I'm afraid.

Yes! I absolutely have! And I've been thinking a lot lately about how to prevent them from collapsing into this state. It is getting better with newer and better models though. But yes, there comes a point when the codebase size and complexity crosses a threshold and we enter a churn state with very little progress.

I was motivated to make the codebase visualizer linked above to help me understand how the various pieces of the system that is being vibe coded connect with each other.

Now I'm thinking about how I can build something similar for agents. How can we teach "systems thinking" to a model?

With every new model I've been attempting to build something that takes at least 1 month and really pushes the model. So far I've done:

- Python interpreter in Rust (fail): https://github.com/BlueBlazin/pyrs

- Low-poly Google Earth like engine (success): https://github.com/BlueBlazin/lowpolyearth

- Codebase -> 2d infinite-zoom visualization (success): https://github.com/BlueBlazin/blockviz

I have never understood or liked all the obsession with one-shot builds.

The fediverse sucks and only people in their small tech bubbles don't seem to realize that the vast majority of people do not know or care about things like "federated social media".

We just need someone* to create a "WoW Classic" version of a social network so that we can get back on it and this time try to make it work. Social media is one of the worst things that has happened to humanity since the web was created. But it used to be good. So it's incomprehensible to me why no one has made a social networking site matching old facebook and without the awfulness added later.

* Maybe some of us can do it, send me a message.

How is this even possible? The change is small and literally has an ethereum URL in the code changes. How did it go through?

I made a stylized low-poly 3D Earth engine using open maps and elevation data that runs in the browser (desktop only for now).

It's written in Rust and runs on WebGPU (wasm with threads). I made it as a native mac app but have now ported it to wasm.

Disclaimer: The app was fully vibe-coded over a course of around 80 days using Claude Code and Codex.

Web app: play.lowpolyearth.online

Press esc for controls.

Home page with more screenshots: lowpolyearth.online

Thanks! I am trying to port it to the browser. Keeping those Anthropic GPUs and TPUs busy.

Thanks, this is an online app and it uses protomaps[1] and mapterhorn[2] data for cities and elevation. So the app streams relevant elevation tiles based on your current location and uses those to render the terrain.

[1] https://protomaps.com [2] https://mapterhorn.com/

I think you may be confusing autoregressive and generative. But that can just mean it's a BERT-like encoder model that outputs a single vector for the given input tokens. To make it sound even more "novel" you can say it's a bidirectional encoder-only transformer model which modern LLMs are not and sell that as something revolutionary.

I really appreciate this type of articles. I feel like a lot of knowledge in LLM training and inference is locked inside the heads of practitioners. Similar to compiler engineers before.

To work in LLM training/inference you’re expected to know this stuff but to know this stuff you need to be working in the space.

Not difficult, just boring.

As someone with a bit of interest in programming languages (design and implementation) this was really interesting to me:

> For most people who work on programming languages, the easy part is coming up with new and better ideas about how to make programming better. The hard part is convincing anyone to actually use those ideas for real work.

Totally agree, there's only so much strangeness you can introduce in a new language[1] regardless of benefit.

But AI agents should not feel much resistance to radically new ideas in PL design. I've been thinking for a while now about how PL design will evolve post agentic AI. I think it will be very interesting to see what new ideas we can come up with to improve programming languages when we worry much less about adoption.

[1] https://steveklabnik.com/writing/the-language-strangeness-bu...

I think they poison outputs now if they detect distillation attempts. So a model trained on distilled outputs will be stupider.

Thanks for the kind words! Even though this is definitely meant for desktop screens, I really didn't want it to fall apart on mobile. As for upvotes, it's super niche so I wasn't really expecting even a single one tbh

A fully vibe coded python 3.14 interpreter in Rust: https://blueblazin.github.io/pyrs

Now at 350k lines. Native and wasm binaries (you can try the limited wasm version online). Currently adding a full CPython test suite benchmark.

Just for fun, not trying to replace CPython here. Mainly to test the limits of current coding agents.

As I understand Peter had already early retired because of a successful startup exit and presumably has more money than he knows what to do with. Does that help make you feel less jealous on him getting a job at oai?