teaearlgraycold

5 ms·

1.
by teaearlgraycold 5d ago
Never had any issues using it for personal small scale development. I think the premise of OpenRouter as a magical way to fall back on error or select providers based on price/tok-per-sec/uptime isn’t delivered upon. But it’s an e
2.
by teaearlgraycold 6d ago
Glad I have my Forgejo service set to auto update.
3.
by teaearlgraycold 7d ago
Phones all have hardware security modules with keys stored inside. That will absolutely be in use for this new feature.
4.
by teaearlgraycold 7d ago
Opus 5 writes too many comments. Other than that I don't agree with what I'm seeing online. It works great.
5.
by teaearlgraycold 8d ago
I want them to be sterile and inhuman as much as you do. But I don't draw the line at "I". I'd rather not read through even more awkward English as it tries to work around how all of the training data has something or so
6.
by teaearlgraycold 8d ago
The AI tattoo. Hope you don’t have any ragrets.
7.
by teaearlgraycold 8d ago
I thought the wisdom is more so don’t bother going below 4bit and you won’t see a difference above 8bit.
8.
by teaearlgraycold 8d ago
Well I definitely don’t like coding as much as you do. And don’t give up your hobbies to appease anyone, least of all me.
9.
by teaearlgraycold 9d ago
They don’t hand write it because it would be an embarrassment. At the same time, Claude Code doesn’t really need to be any better than it is while Anthropic has great models no one else has.
10.
by teaearlgraycold 9d ago
I have some lines of code I’ve written that I quite cherish. I’ve posted about such excerpts many times online. So we both like code quite a bit. I think you’re letting your pride get the best of you. Writing software should not be about eg
11.
by teaearlgraycold 9d ago
I did. Still missing the Vulkan support for Zed. But I noticed my WebGL ray marched shaders run quickly - with glitchy artifacts.
12.
by teaearlgraycold 10d ago
All of that IP stolen by OpenAI and they didn’t think to buy some goodwill lifting a copy of an engineer’s manual to their GPUs…
13.
by teaearlgraycold 10d ago
Depends on your desktop UI. If you install something like Gnome (as I did with Asahi Linux today on an M3 Air) the animated transitions don’t work. You just get a short frozen period and then it cuts straight to the final state. I could try
14.
by teaearlgraycold 10d ago
It’s cooler to run Linux on bare metal
15.
by teaearlgraycold 11d ago
I don’t see how you could run Qwen3.8 27B on 16GB of memory that’s shared with Linux. Are people running models at 2bit quants? Are they even worth bothering with? I had assumed you go down to 4bit and if you need to go smaller you have to
16.
by teaearlgraycold 11d ago
Charitably interpreted, GP is already loved and takes it for granted while arguing about a straw man. Or perhaps GP is specifically talking about someone who has trouble loving themself without a romantic partner.
17.
by teaearlgraycold 11d ago
As someone in my 30s I'm not sure I need the things on your list. Can't escape the aging of my parents but I'm lucky they will be able to financially support themselves through retirement. If I do add complications to my life
18.
by teaearlgraycold 12d ago
I’m confused why he wasn’t scripting the deletion process.
19.
by teaearlgraycold 13d ago
I just use Claude Opus and the GLM series. Nothing’s really changed for my workflow in the last 6 months.
20.
by teaearlgraycold 13d ago
I wish they'd turn it into a long thin park. Keep just enough space for buses and maybe cut into the sidewalks a bit for greenery.
21.
by teaearlgraycold 13d ago
Fair enough. I didn’t do most of the work.
22.
by teaearlgraycold 13d ago
I'm not producing commits. LLMs are producing commits. Fix more things, solve more problems, and do it to a higher degree of quality in less time using LLMs. What's to complain about?
23.
by teaearlgraycold 13d ago
Coverage started out kind of mediocre, about 50% IIRC. There were a lot of easy wins, and a handful of files were intentionally ignored. I also had it first review all tests for superfluous ones that could be deleted.
24.
by teaearlgraycold 14d ago
On a technical level it would still be a nightmare. But yes it would actually get done.
25.
by teaearlgraycold 14d ago
> Constantly having someone looking over your shoulder is a world of stress and overstimulation that I (and I suspect many others) will not stand for. I suspect this is a simple social anxiety you could train away.
26.
by teaearlgraycold 14d ago
I said inflated. They would always be expensive. But they charge more than their competitors per flop and per byte because they’re the only manufacturer that can run CUDA. If they lost that moat their prices would go down a notch.
27.
by teaearlgraycold 14d ago
Not sure what you’re getting at
28.
by teaearlgraycold 14d ago
Doesn't really mean anything without a specific use case to guide model selection. The thing is Qwen 3.8 27B can be ran on far far cheaper hardware. If you're spending the big bucks on these rigs you probably made the wrong choi
29.
by teaearlgraycold 14d ago
These GPUs are extremely inflated in price because Nvidia effectively has a monopoly on hardware that is used to train models. Apple Silicon tends to have good inference software available but as soon as you want to train even a YOLO model
30.
by teaearlgraycold 14d ago
All sorts of efforts that used to be put off indefinitely can now be handled largely by LLMs. Yesterday I took a codebase (~50 source files) and spawned a sub-agent (GLM 5.3 Flash) for every single file. Each file was analyzed for test cove

More ›