behaviors

This is very SaaS optimal. I never use these types of services on a "smartphone" device. I honestly think my OS (derived from AOSP) is more secure than googles "trust." Every android device is not an identity tied to a account. Many phones exist with no accounts whatsoever.

Yeah the next generation is understanding the codebase vs understanding how to prompt the info out of the codebase. I too took on the rust port of a working software, learned some LLM roadbumps, and it looks very much like the future is LLM coded and human "managed"

Singapore 5.53

Is this chart in claudes per hour? The graphs have bad indexes/legends. I assume this is at the top because of population density? Overall, it's not that useful for me, but I'm positive it's driving Anthropic's roadmap.

My state has 1% of claude users writing code.

It is a far bit tougher to actually get the clankers to speak accurately. I understand the legal perspective, with OpenAI talking about depression use cases, these companies who are running computers for users have to worry the software might harm the user(through themselves) and the leagl fallout needs protected.

It amazes me that we are going to litigate this like they did with cars over horses, or machines vs human labor. I honestly don't think Claude should be running companies.

I used the spaghetti-detective plugin/add-on for OctoPi when I got my printer, they also hit bandwidth of video streaming over web(part of the "monitoring" area) they seemingly have been absorbed into "obico"(the github remains github.com/TheSpaghettiDetective) Every 3DPrinter software has options to replace these Bambu Cloud features, the process involves a fair bit of deep dive understanding, flashing firmwares, troubleshooting bugs, and then you could in theory use the same machines with all the Bambu Cloud features, in a local environment.

My only gripe with the community approach is, why not replace them rather than attempt to use ANY servers they have? Jeff cleverly highlighted that all the slicers originate from Slic3r, there is always a point before Bambu.

A model framework for an in house suite of models.

From dataset harvest, to training intricacies on CUDA/ROCm to fun HIP kernels. Full circle to inference testing, building it around consumer hardware(the challenge). Using this as a "how it works" deep dive, allowing me to learn more about the how, more than endless papers will. It's a MoE and I'm slowly running a human loop, research, build, correct, research.

If the atomic action is idempotent, you don't need a layer for repeating yourself. You hit the nail on the head. So much idempotency efforts are made because they never made the actions idempotent in the first place.

Way to roll the nostalgia. AIM and rotten, seeing grotesque human sacrifice and torture at "13" was a unique time to be alive.

It's a HTML browser for Pi Pico2, CLI, meant to support my in-house project running on a mesh of Pico2's. I really wanted to use RISC-V and it needed a webhook that serves a page on PIO wake. We are at the browser is written about 60%. The server is now already handled ;P I found this awesome project someone posted on HN. When I complete my project the browser will be released alongside. You can very likely reproduce it with less than a handful of prompts. One thing I really do believe, ideas are going to be the next open-source. LLM's can make ideas into things.

I do believe this is just a next step in languages. We've come this far trying to make code NLP, now we have the closest thing to a translator in our generation. It's an exciting time, just don't pay attention to talking heads.

Well done. Been working on a similar smaller project for RISC-V. This is excellent

As a former tie a new knot/day challenge taker. I really like this! Making knots into a formula is awesome. That paper is a great read in itself. Fun was had, that random gen with ikea images is top notch!

It's funny, because that task is very diverse. Any LLM will use the codebase given as a template(At least in free-tier models)

My software as a contract of behaviors works like a program bench(I even cross tested buildouts) Made an entire corpus layout for multi agent multi platform builds to be compared. Even went ahead and ran 50 contracts for an example. It honestly showed improvable areas, and distinct differences between model code.

{contract_name}/ └── submissions/ └── {date}_{os}_{agent}_{model}_{stack}/ ├── {contract}.osc.md ├── osc.osc.md └── results/ └── {contract}.snapshot.json That's it, compare to the same contract, or find a new contract to use to compare. Lot's of signed/hash pinned files are all you need to reproduce software from nothing, with an LLM.

Programbench is close to that(they have a nice paper/article here. But I don't like the work used. Having software to start with is not a bench of making code but reverse engineering.

github/s1ugh34d/osc

I've been using an OpenSCAD container with various local models. Dumping the render.png straight to the model, allowing it to modify the code and try again. Made some interesting things, but the main purpose was to fix things I've already made and have some weird single issue that cascades to a broken model if I touch it. OpenSCAD is the first step, FreeCAD and similar(now starting to see more CAD LLM work) are still a WIP. Since january we've solved 4 solid issues I've left on backburner. I use the docker container version with some Custom wrap/bridge work for the render dumps.

Most of the big hit's in tech had a trendy index swinging moment, Intel has been searching for one for a long time since AMD64 undercut the Itanium. Hype drives a currently multi-billion dollar bubble. It's not always a bad idea to throw our holy noodles at the wall. You might find they hover is the sky and grow meatballs, could be big.

The exact reason you should start with one first. I support maximum specification. Atomic if you will. Lucky for modern development, you don't have to write it, you just have to proof it. If you can read a spec, it can guide your development, might as well have a system to manage them.

It's almost once a week I see a very similar approach, with variables for the developer's workflow.

Sounds just like what I've been using, and published as a template: github.com/s1ugh34d/osc (It has a results corpus of testing done) Specs that are LLM driven development, behavior addressed. All markdown human written, LLM's can write them too. The result is tight spec outcomes from LLM's assisting workflows.

Using massive human software specs was great with human developers, but with "clankers" the problem is context. How to wrap software contract context into a markdown file. Works for me, I can have Claude make a osc, review it, amend it, and build a simple solution that has verifiable outcomes.

This SpecDD is very similar, doesn't really need to book of documentation, just a guide really.