g023

5 ms·

1.
by g023 8d ago
They need to make a Big Bonsai, something at the enterprise levels that can compete with DSV4 Flash etc.
2.
by g023 9d ago
While the North American models say 'No', the chinese models say 'Go Go Go'. I guess we'll see whether the anti-consumer wins over the pro-consumer.
3.
by g023 11d ago
Working on a DeepSeek V4.x coding harness. Added ability for the llm to set its own reasoning for each response, bi-directional communication between subagents, added a todo sub orchestrator that for each task a subagent does, it passes its
4.
by g023 16d ago
And you don't think it is weird that the chinese models are the only ones embracing free markets, and open concepts, while the so called 'free' market on this side of the world tries to close everyone off from the technology
5.
by g023 17d ago
its probably more expensive to run, and I'm thinking 4.1 flash is a smaller more efficient model. You can always host your own. This is what they need to do to stay competitive.
6.
by g023 26d ago
I miss the old days of ICQ and just dragging a file onto the person you are sending your file and bam, done like dinner.
7.
by g023 1mo ago
All this performance at such small model sizes, why are the API fees so high for the AI monopolists on this side of the world?
8.
by g023 2mo ago
g023 Code - https://github.com/g023/g023code a deepseek v4 flash harness built specifically for that model, that uses Ollama for its vision support, and uses the built in web search capability natively. Created only in
9.
by g023 2mo ago
It says 'yes' where the others say 'no'. Good enough for me.
10.
by g023 2mo ago
Pure-Python AI coding agent powered by DeepSeek V4 Flash - Subagent-First architecture · Context is Currency · Terminal-native - Can connect Ollama models for Vision support - Uses the native web search built into DS v4 API and is built aro
11.
by g023 3mo ago
Slap the gpus in a car and offset the cost of ownership by supplying the grid for GPU power on the go. Either get paid in rebates or tokens. Contribute to a distributed training/inferencing network.
12.
by g023 3mo ago
Anything to close Pandora's box. "They" liked the eras they could control the communications, and therefore the narrative. Boomers on their last legs, question is, will the future undo the unjustness that was forced upon them
13.
by g023 3mo ago
gee I wonder how their models learned Chinese?
14.
by g023 3mo ago
How come people these days treat jobs like its a social gathering?
15.
by g023 4mo ago
A self-contained CUDA inference engine for LiquidAI/LFM2.5-8B-A1B (hybrid conv + GQA-attention MoE, 8.5B params, 1B active) targeting a single RTX 3060 (12 GB) using flash-decoding. MIT license.
16.
by g023 4mo ago
I wonder if "battelites" might be profitable. Like an pay-per-usage energy grid in space with battery backup that can beam power around to other satellites that might not have easy access to power, or have their power grids tempor
17.
by g023 4mo ago
China has a couple going at the moment.
18.
by g023 4mo ago
When did "hate the customer" become a thing?
19.
by g023 4mo ago
I use DeepSeek v4 flash with CoPilot and it works pretty good.
20.
by g023 4mo ago
If anyone is looking to hook it up to copilot, I made a proxy script to handle the connection a bit back that might be handy: https://gist.github.com/g023/c2bb7b540ffe64cee76023f18f6f936...
21.
by g023 5mo ago
Terminal-based chat application powered by *locally installed* llama.cpp, featuring an auto-managed server backend, reasoning modes, and 7 built-in filesystem tools for interactive AI assistance with a focus on only allowing read only agent
22.
by g023 5mo ago
We need more personal level AI solutions instead of so much corporate centered solutions.
23.
Local Model Router: Ollama/OpenAI-compat bridges for local LLMs via llama.cpp
1 points by g023 5mo ago|0 comments
24.
by g023 5mo ago
I've started creating https://github.com/g023/localmodelrouter/ which offers Ollama like functionality but as a single .py file with minimal dependencies and more focus on letting llama.cpp handle the dirty w
25.
by g023 5mo ago
HarnessHarvester generates executable Python harnesses from natural language task descriptions, executes them in a sandboxed environment, reviews them with multi-faceted LLM judges, and repairs failures using branching strategies. It includ
26.
by g023 6mo ago
A single file, python based, minimal/recognizable dependencies, turboquant playground, barebones af, with some easy to access globals to experiment with at top of 'run_tquant.py'. Test model is a 1.77B model that I altered by
27.
by g023 6mo ago
I had some issues in the original, but had to jump away for a bit here to do some backups (weak). Anyways, I updated to make the necessary fixes, and also made some more tweaking values at top to play with and dialed in the params for the m
28.
Show HN: Standalone TurboQuant KV Cache Inference(github.com)
3 points by g023 6mo ago|4 comments
29.
Show HN: An offline first focused agentic CLI application powered by Ollama
4 points by g023 6mo ago|0 comments
30.
Show HN: G023's Agentic Chat with Memory and Python Power(github.com)
1 points by g023 6mo ago|1 comments

More ›