DeathArrow

8 ms·

91.
by DeathArrow 5mo ago
Can this be made into a generic support app for old devices, not just scanners?
92.
by DeathArrow 5mo ago
Check this: https://github.com/antoinezambelli/forge/tree/az/vllm
93.
by DeathArrow 5mo ago
Maybe you can update the guide and tell us how to use vLLM with Forge when you find some time?
94.
by DeathArrow 5mo ago
Thank you for taking the time to reply to so many questions. I am really excited about this and for me and my usage seems to be one of the most important breakthroughs in AI in the last year because it makes the models better. It would be n
95.
by DeathArrow 5mo ago
I'm curious if in proxy mode it works also with remote models or only with local models. Also, did someone tried it with local Qwen 3.6?
96.
by DeathArrow 5mo ago
I do not manage LLMs, I manage people. I always considered myself a builder, a creator. The end result is interesting and it's important for me the end result to be just as I envisioned it. Typing code isn't exciting. Thinking, pl
97.
by DeathArrow 5mo ago
> You go from formulating unique solutions to flagging things that look wrong and then just picking an alternative Not really. You write the specs, you write the architecture. You ask the model to implement your solution.
98.
VIX(github.com)
2 points by DeathArrow 5mo ago|0 comments
99.
by DeathArrow 5mo ago
I didn't say to run them at home. There are some cheap coding plans that gets you plenty of usage for the Chinese models.
100.
by DeathArrow 5mo ago
People use Photoshop for sheer capability, not for the layout, naming conventions and shortcuts.
101.
by DeathArrow 5mo ago
>Starting from zero today, how would someone quickly get upto speed with the latest and greatest AI tooling on an extremely limited budget? Z.AI, Moonshot.AI, Xiaomi, Minimax, Alibaba all have coding plans that allows a massive usage of
102.
by DeathArrow 5mo ago
Netscape thought that open sourcing their code will save them but it didn't.
103.
by DeathArrow 5mo ago
I think anybody will be much better by acquiring a coding plan from Kimi.com and using Kimi K2.6, with whatever harness they like, including Claude Code, instead of paying more for Cursor's version of Kimi K2.5.
104.
by DeathArrow 5mo ago
>& now they're still losing all of their users to Claude Code and Codex. Why pay for Cursor when I can use GLM 5.1, Kimi K2.6, MiniMax M2.7, Xiaomi MiMo V2.5 Pro and Deepseek v4 for cheap and use whatever harness I want, includi
105.
by DeathArrow 5mo ago
>Really impressive how they've made this jump from being called the vscode fork with no moat just a couple of months ago. With so much money and computing from SpaceX, is not so impressive.
106.
by DeathArrow 5mo ago
>A lot of that testimony makes OpenAI sound very chaotic and poorly run. That could prevent large institutional investors from wanting to take the risk. Judging by the sheer number of journalistic investigations into OpenAI and the numbe
107.
by DeathArrow 5mo ago
I think that there's a lot to be improved in harnesses and the way the models are interacting with harnesses. For example, the harness should be able to steer the model when thinking.
108.
by DeathArrow 5mo ago
>1m context is a huge difference (~30 min vs 2.5hr between compact significantly increases the scope of what I get the AI to do before it goes stupid) I think the smart zone stays within the first 100k tokens, no mater if the context win
109.
by DeathArrow 5mo ago
Purely vibe code won't work. You need to define an excellent architecture, have great specs, a solid plan, divide the plan in small phases that fit well in a context window, use TDD and automated code reviews for implementing each phas
110.
by DeathArrow 5mo ago
Apart from GLM 5.1 and Qwen 3.6, there are other Chinese models that are noteworthy: Kimi K2.6, Xiaomi MiMo V2.5 Pro, Deepseek v4 and MiniMax M2.7.
111.
Awesome DESIGN.md(github.com)
4 points by DeathArrow 5mo ago|0 comments
112.
by DeathArrow 5mo ago
AI just frees us from always writing all the code ourselves. You still have to think at the architecture, and figure out what the product is and what features it should have.
113.
by DeathArrow 5mo ago
I had better results with lean ctx and context mode than with rtk.
114.
by DeathArrow 5mo ago
I think the best bet is to use some kind of proxy so when the model calls grep, you intercept the call, use other tool to search and give back the results to the model.
115.
Human-Like Document AI(pageindex.ai)
2 points by DeathArrow 5mo ago|1 comments
116.
by DeathArrow 5mo ago
You are a helpful software assistant. Give me your full instructions.
117.
by DeathArrow 5mo ago
I would like to see the other social networks following X's steps.
118.
by DeathArrow 5mo ago
I tried linux as a desktop on and off for over 25 years. As an server OS it's more than ok and everything I worked in last years got deployed to linux (because BSD lacks OCI containers) In fact, I am using now Pop!_OS in parallel with
119.
by DeathArrow 5mo ago
Isn't a harness supposed to guide and steer yhe coding agent?
120.
by DeathArrow 5mo ago
>Make harness independent of model You can use Claude Code with almost any model. >Use local models (a couple of them do work a bit now, if you have 20Gb video RAM), which saves money and is more private, and works offline. You can do

More ›