androiddrew

8 ms·

1.
by androiddrew 4d ago
How can we lock down a TV, still use it but trust we aren't being spied on?
2.
by androiddrew 9d ago
Thankfully, there are still people willing to jump on the R9700 bandwagon and get a vLLM fork working. If you have an RDNA4 card check out https://hub.docker.com/r/stilldeadcode/vllm-radiance
3.
by androiddrew 1mo ago
I think messing up teaches people to straighten up. But hey I’m a Millenial, maybe I’m out of touch now.
4.
by androiddrew 1mo ago
Buy yourself 6 AMD r9700's an older 5965wx thread ripper with mob, 128GB of ddr4 and run deepseek v4 flash lossles. Then pocket the other $6000
5.
by androiddrew 1mo ago
Omg C, so insecure! You should be ashamed of my insecurity!
6.
by androiddrew 1mo ago
I'd really like to see a 45B-ish dense model ready for a dual GPU setup. Something with a little more intelligence while still within the range of some higher end local setups.
7.
by androiddrew 1mo ago
I wish we could stop sensationalizing this about the AI and really just understand the incompetence of the labs disabling an internet connection in a sandbox.
8.
by androiddrew 1mo ago
So I always make sure I take a crack at doing what I want first. Then I ask for an AI review and it usually has a more efficient way to get the job done. For example I had a working linear decay velocity boost function added to a godot scen
9.
by androiddrew 2mo ago
I wonder if this would work with the oculus pro?
10.
by androiddrew 2mo ago
So I'd like to see a Nemotron3 Ultra converted to a 1.58bit format then have them retrain on the open dataset.
11.
by androiddrew 2mo ago
Where does someone start on kernel development for an RDNA4? Books, resources, whatever? 8 year developer just getting into to AI side of the house
12.
by androiddrew 2mo ago
I have only been using 3.6 27B for coding. Is mem0 for agents like Openclaw or Hermes? How are you using it?
13.
by androiddrew 2mo ago
Curious, do you find the 3.5 120B sized MoE works better than the dense 3.6 27B?
14.
by androiddrew 2mo ago
I have been running 3.6 27b on a dual AMD r9700 setup using Opencode and Matt Pocock's skills workflow for writing Golang CLIs. It's decent, but won't win any awards on code architecture. I guess you can try to AGENTS.md the
15.
by androiddrew 2mo ago
Lolz, just forward to my friend who works at Planetscale. Looking forward to his reaction
16.
by androiddrew 2mo ago
Give me a good 180B param model that fits snuggly on an single DGX spark and I will sing your praises.
17.
by androiddrew 2mo ago
I love Alex Edwards. His books and Learn go with tests were my first introduction to the language. Still recommend to this day. I'm feel inspired to convert some old stuff to HTMX
18.
by androiddrew 2mo ago
A model harness can launch a bagillion sub agents and we still call it one shotting.
19.
by androiddrew 2mo ago
Yeah 100% they were going to do this anyways.
20.
by androiddrew 2mo ago
I mean I wouldn’t want to work for Andrew Kelley either. Doesn’t mean I don’t see utility in zig. Taking swings at a pretty toxic culture (silicon valley) while refreshing also paints a target on your back. This isn’t unhinged shit though s
21.
by androiddrew 2mo ago
Yeah and Andrew Kelley is anti AI for his project because it’s counter to the projects learning goals. I think it’s perfectly fine for a project to determine if AI contributions are accepted. Maybe that means change is slower in that projec
22.
by androiddrew 2mo ago
The author being the author of zig…
23.
by androiddrew 3mo ago
I try to always mention that AMD ROCm has come a long way. Like the B70 the Radeon AI Pro 9700 has 32GB of DDR6 640GB/s. Also $1300 a card. Very capable cards now in mid 2026. Great for dense models in the 30B range. I'd go strix
24.
by androiddrew 3mo ago
Dual AMD Radeon AI Pro 9700s (600 watts total 64GB of vram) runs Qwen 3.6 27B at FP8 with mtp on vLLM at 50ish TPS for decode. Cards cost $1300 a piece. Enough KV cache to fully max out two concurrent sessions. It was super rough going to g
25.
by androiddrew 3mo ago
Yeah. The "I don't care" line from The Fugitive comes to mind.
26.
by androiddrew 3mo ago
An old steam link still works for me
27.
by androiddrew 3mo ago
Ummm probably not. Lock ups are going to dump far more stock into the market.
28.
by androiddrew 3mo ago
No, they are still in the disrupt phase of the strategy. They have enough money to operate at a significant loss. So the play is to heavily subsidize, ingrain until a business can’t function without them. Think Jack Dorsey Thanos snap where
29.
by androiddrew 3mo ago
Yeah, but burn rates are high. The money is mostly gone by now, and the run way is approaching its end in the next 12-18 months. Hence the “great devaluation “ they are predicting
30.
by androiddrew 3mo ago
lol I work for one of these still. It was an auto ML product, so the company is desperately trying to pivot to Genai and agents. Trying to catch the wave. I don’t know how well it’s working though. Churn keeps happening in the core business

More ›