mythz

13 ms·

211.
Generate Blazor Admin CRUD Apps from a Text Prompt(servicestack.net)
1 points by mythz 2y ago|0 comments
212.
Generate Blazor Admin CRUD Apps with Text to Blazor(servicestack.net)
1 points by mythz 2y ago|0 comments
213.
by mythz 2y ago
Working on a self-hosted OSS AI Server with support for LLM APIs (OpenRouter/OpenAi/Anthropic/Google/etc), Ollama endpoints, ComfyUI and FFmpeg agents. It supports Synchronous, Queued and Reply to Web Callback APIs for e
214.
by mythz 2y ago
Looking forward to it! I'll generally make an effort to use Open Models over proprietary alternatives when the use-case permits as Open Models getting better and more popular encourages more models to become open as well - a requisite
215.
by mythz 2y ago
I've started using SQLite for all new Apps a couple of years ago and haven't looked back. It was the catalyst for us to get off AWS (and RDS) where we now deploy (.NET) Docker Apps with Kamal to a shared Hetzner Linux VM which use
216.
by mythz 2y ago
As a longtime Windows user I switched to a Fedora Desktop. Prefer using a modern distro with latest software inc. Wayland and latest Nvidia drivers and Flatpacks. No idea what's possessed Ubuntu to go their own way with Snap's, bu
217.
by mythz 2y ago
First I've heard about issue with bash. I've been using OhMyZsh in all terminals for years which seems pretty active and auto updates fine whereever I'm using it, e.g in Linux, macOS, Windows/WSL. Zsh might have a slower
218.
by mythz 2y ago
Happy to seem them back to releasing OSS models, we used a lot of their OSS models early last year before they were eclipsed by better models and never bothered to try any of their large models which IMO weren't great value.
219.
by mythz 2y ago
Didn't expect to be cheering for Chinese AI companies and Facebook over mega funded US tech corps, but here we are. Were fortunate that not all SOTA AI models are controlled by US Tech corps. Right now they're in the "maximum
220.
by mythz 2y ago
They can be linked, e.g. 2x DIGITS can run 405B models [1]. Won't know what value/performance we can get until they start shipping them in May. https://nvidianews.nvidia.com/news/nvidia-puts-grace-blackwe...
221.
by mythz 2y ago
Someone also got the full Q8 R1 running on a $6K PC without a GPU on 2x EPYC with 768GB DDR5 RAM running at 6-8 tok/s [1]. Will be interesting to see the value/performance compared to next gen M4 Ultra's (or Extreme?) vs NVID
222.
by mythz 2y ago
> i absolutely condemn them and others (e.g Facebook) for their deliberate and continued misuse of the term This is the kind of inconsequential nitpicking diatribe I'm referring to. When has "open data" ever meant Open Sou
223.
by mythz 2y ago
This nitpicking is pointless. DeepSeek's gifts to the world of its open weights, public research and OSS code of its SOTA models are all any reasonable person should expect given no organization is going to release their dataset and op
224.
by mythz 2y ago
It's one of the things I love about Bun, I now write any non-trivial stand-alone scripts with TypeScript which is capable of a lot of functionality without needing any config or to install any external packages.
225.
by mythz 2y ago
No kidding, I thought global tools worked with fairy cloud dust, now I wont be able to sleep! But knowing there's not a going to be a bloated node_modules folder and local dependencies running different versions unnecessarily maintaine
226.
by mythz 2y ago
From its documentation [1] it looks a lot like a parceljs replacement [2], i.e. a zero config bundler which processes and bundles the dependencies in .html pages. So great for simple websites, not for replacing an entire Vite stack. [1] ht
227.
by mythz 2y ago
Couldn't think of a project that was more doomed to fail than a competing alternative to Node.js, but glad that I gave it a shot when I needed to create lots of stand-alone scripts to process text files and SQLite DB updates which I wa
228.
by mythz 2y ago
I've linked to why. We prefer #NoBuild solutions where we'd only use local npm deps/node_modules if we absolutely have to, and for non JS Apps we don't.
229.
by mythz 2y ago
We don't, we use npx to avoid needing any local dependencies. `npx tailwind@v3` is the solution, definitely wont be committing node_modules to git, ever.
230.
by mythz 2y ago
Yep `npx tailwindcss@v3` is the solution we ended up with. Just had to change this in all our Apps and project templates.
231.
by mythz 2y ago
We only use npm dependencies and bloated node_modules folders [1] if we absolutely have to and for our non node.js Apps we explicitly don't. [1] https://www.reddit.com/r/ProgrammerHumor/comments/6s0wov&#x
232.
by mythz 2y ago
FFS we don't have local dependencies. Which means we don't have any local node_modules and have never ran npm install (there are no dependencies to install!). There is no lock file or dependencies. Our Apps don't have any npm
233.
by mythz 2y ago
Not everything is a node App. We don't have a lock file because we don't have any local npm dependencies (as already mentioned). We're using `npx tailwindcss` as-is.
234.
by mythz 2y ago
Was it really necessary to break all existing apps using `npx @tailwindcss`? It's an easy enough change to `npx @tailwindcss/cli` but we now have to go back and update all Apps and templates and since it uses v4 we now have to tes
235.
by mythz 2y ago
Works great for us as most of our code is public and we can only benefit from more our code of our product or using it being available. Also happy for any of our code expands their training set and improves their models even further given t
236.
by mythz 2y ago
The Mixtral grading model calculates the original starting votes which can be further influenced by Users voting on their preferred answer which affects the leaderboard standings. It should be noted that Mixtral 8x7B didn't grade its o
237.
by mythz 2y ago
The software that controls the front-end has enormous value, it becomes the central point and brand to manage and self-host LLMs that's used to manage 100 GB catalog of models which acts like a moat inhibiting switching to alternatives
238.
by mythz 2y ago
Yeah we evaluated several models for grading ~1 year ago and concluded Mixtral was the best choice for us, as it was the best model yielding the best results that we could self-host and distribute the load of grading 1.2M+ answers over seve
239.
by mythz 2y ago
Was disappointed in all the Phi models before this, whose benchmark results scored way better than it worked in practice, but I've been really impressed with how good Phi-4 is at just 14B. We've run it against the top 1000 most po
240.
by mythz 2y ago
Calling someone selfish because he chose to devote his efforts to one of his spinoff projects and not wherever you think he should devote his future time/efforts to goes beyond entitlement. It's even worse coming from an anonymous

More ›