mythz

11 ms·

181.
by mythz 1y ago
Because the pricing when creating an instance shows me the cost for the entire month, then works out the average hourly price based on that. This is just creating a GPU VM instance, I don't see how to see the cost of different NVidia G
182.
by mythz 1y ago
Just the Developer plan, haven't hit its limit yet.
183.
by mythz 1y ago
Yep it's off to a great start with its early mover advantage but IMO their days on top are numbered with every major player behind the premier coding models (and major IDE vendors) iterating hard on their own integrated AI coding agent
184.
by mythz 1y ago
IMO it's a strategic misstep to try and create their own IDE with a fork of VS Code. I'm only going to consider AI Tools that integrate with my IDEs (primarily VS Code + Rider) as such my AI weapons of choice are now: augmentcode.
185.
by mythz 1y ago
Nice, only got the pro plan and didn't want to get Nickel and Dimed per usage so ended up ignoring claude code, but will give it another go now to see how far I can get with the Pro plan. Turned out to be a blessing as it led me to dis
186.
by mythz 1y ago
RunPod also charges per second [1], also this is Google's expected avg cost per hour after running it 24/7 for an entire month, I couldn't find an hourly cost for each GPU. When you need under <1hr than you can go with Run
187.
by mythz 1y ago
The pricing doesn't look that compelling, here are the hourly rate comparisons vs runpod.io vs vast.ai: 1x L4 24GB: google: $0.71; runpod.io: $0.43, spot: $0.22 4x L4 24GB: google: $4.00; runpod.io: $1.72, spot: $0.
188.
by mythz 1y ago
Awesome stuff, this resolves my #1 feature request of being able to replicate an entire directory of SQLite *.db's from a single Litestream process - happy it's finally here. Should make replicating Multi tenant per-user SQLite da
189.
by mythz 1y ago
China doesn't know what you're talking about.
190.
by mythz 1y ago
Lots of words to say OpenAI will remain an SABC (Sam Altman Benefit Corporation)
191.
by mythz 1y ago
Abandoning Windows after 25 years for Fedora gets validated with every "Windows Innovation"
192.
Open-Source AI Server – April 2025 Update(servicestack.net)
3 points by mythz 1y ago|0 comments
193.
AI Server – April 2025 Update(servicestack.net)
1 points by mythz 1y ago|0 comments
194.
Open-Source AI Server – April 2025 Update(servicestack.net)
1 points by mythz 1y ago|0 comments
195.
by mythz 2y ago
The speed gains are real, after downloading latest QAT gemma3:27b eval perf is now 1.47x faster on ollama, up from 13.72 to 20.11 tok/s (on A4000's).
196.
Fedora Linux 42 Released(fedoramagazine.org)
12 points by mythz 2y ago|0 comments
197.
by mythz 2y ago
On the face of it cutting huge IT govt contracts sounds like a good choice to reduce waste, but with this administration I'm expecting any savings to be funneled to Grok/xAI/Elon.
198.
by mythz 2y ago
Took a bit of research and trial and error to get to a solution we liked, but as it's just deploying and managing Docker Apps, deploying .NET Docker Apps was ok but a bit of time was spent supporting DB Migrations and litestream and ru
199.
by mythz 2y ago
Will still be using https://kamal-deploy.org to deploy my .NET Docker Apps to https://cloud.hetzner.com
200.
by mythz 2y ago
All these user hostile behavior finally tipped me over the edge last year after I started to get Windows 10 EOL popups. No way I was going to "upgrade" to a ad/spyware/subscription ridden MS marketing platform and be con
201.
by mythz 2y ago
That’s not a flex, it requires no skill to install software, they’re just some of the better tools you can install to boost productivity in Linux terminals. I doubt they’re obscure to any Linux CLI user who spent time on improving the defau
202.
by mythz 2y ago
Weird flex, most commands, utilities, server software and remote tools are going to be run are going to be from the command-line. All our System Administration of remote servers uses the command-line as well since exclusively deploying to L
203.
by mythz 2y ago
Gaming/WSL kept me on Windows for a lot of the last decade, however after Windows 10 became EOL'd and Windows started turning into ad/spyware I finally gave it up over a year ago after 25+ years on Windows Desktops. Anyway L
204.
by mythz 2y ago
I understand how developers can come to this conclusion if they're only using local models that can run on consumer GPUs since there's a time cost to prompting and the output is fairly low quality with a higher probability of erro
205.
by mythz 2y ago
Been a happy Fedora Desktop user since switching from 20+ years of Windows last year. Had a few hiccups at the start of last year with NVIDIA drivers in Wayland but that's since been resolved with Explicit sync support. Happy there&#x2
206.
by mythz 2y ago
Concur with Gemma2 being underwhelming, I dismissed it pretty quickly but gemma3:27b is looking pretty good atm. BTW mistral-small:24b is also worth mentioning (IMO best local model) and phi4:14b is also pretty strong for its size. mistral-
207.
by mythz 2y ago
Not sure if anyone else experiences this, but ollama downloads starts off strong but the last few MBs take forever. Finally just finished downloading (gemma3:27b). Requires the latest version of Ollama to use, but now working, getting about
208.
by mythz 2y ago
AMD Ryzen AI Max SoC chips have that [1], but it maxes out at 128GB RAM. [1] https://www.amd.com/en/products/processors/laptop/ryzen/ai-3...
209.
by mythz 2y ago
Yeah VRAM option is good (if it performs well), just sad we'd have to drop 10K to access it tied to a prev gen M3 when they'll likely have M5 by the end of the year. Hard to drop that much cash on an outdated chip.
210.
by mythz 2y ago
Ultra disappointing, they waited 2 years just to push out a single gen bump, even my last year's iPad Pro runs M4.

More ›