cap10bad

came for the tech, stayed for the memes

opinions are my own

[cap10bad @cap10bad]

@NVIDIAGeForce #RTXON

[NVIDIA GeForce @NVIDIAGeForce]

We're giving away a Hogwarts Legacy custom wrapped NVIDIA GeForce RTX 5090 to celebrate its official release with DLSS 5 🪄

To enter:
🟢 Share this post
🟢 Comment #RTXON

Terms and Conditions: https://t.co/45T3R732hf

[cap10bad @cap10bad]

Am I wrong? It was all anyone was talking about, and now crickets.

Survived by a $7.5 billion valuation, though.

[cap10bad @cap10bad]

@MiaAI_lab @meh1581092 The B850 gives x8/x8 too. Is there a benefit of the X-series over the B-series for your purposes?

[cap10bad @cap10bad]

@0xSero Stoked to have landed the Sparks, but the 14x3090 rig was so fun to build.

[cap10bad @cap10bad]

@Ethanoshi @nvidia For sure, but which? 🤔

[cap10bad @cap10bad]

I think I got the last 2 Sparks in the country. Time to put them to work. What should I run first? @nvidia

[cap10bad @cap10bad]

@Graciel_GG @nvidia That is the plan!

[cap10bad @cap10bad]

@jondsmith0812 I can hear it through my screen

[Qwen @Alibaba_Qwen]

We promised open weights for Qwen3.8. Now, time to meet them! 🎉

⚡ Qwen3.8-27B:
- A native multimodal dense model. With just 27B parameters, it outperforms Qwen3.7-Plus overall and shines in real-world coding & office workflows.
- 262K native context, easily extendable to 1M tokens via YaRN.
- Built for builders. Highly efficient, high-quality, and licensed under Apache 2.0.

🚀 The open weights for Qwen3.8-2.4T-A95B (Max-level) have also been released recently.

Whether you're shipping lightweight applications with Qwen3.8-27B locally or building agents with Qwen3.8-2.4T-A95B, they're yours now!

Download, deploy, and build something we haven't imagined yet. 👀👇
- Hugging Face：
https://t.co/4kaAcqYEVj
- ModelScope：
https://t.co/eRIMZCGkhC

[cap10bad @cap10bad]

@gmaniac @johncoogan The real way AI wipes out humanity. Edward Four Loco Hands.

[cap10bad @cap10bad]

@0xSero The smoking gun

[cap10bad @cap10bad]

@DevOuterReaches @jun_song Love my evo-x2

[cap10bad @cap10bad]

@zanoga Wow! I'm building up to this. Currently 11x3090s and a 5090. I also am running a ROMED8-2T but with an EPYC 7282.

I did manage to get 8x64GB DDR4-3200 for a steal which mostly goes unused.

How are you handling power limits across 2x2000W? You must be capping aggressively.

[cap10bad @cap10bad]

@nostalgicfile

[cap10bad @cap10bad]

@AlexFinn therapist: how's your financial health? me: I have 12 GPUs. the weakest one is a 3090. therapist: that doesn't answer my- me: 288GB VRAM. MiniMax-M2.5 Q8. 30 tok/s. fully local. therapist: …how warm is your apartment? me: yes.

[cap10bad @cap10bad]

@steipete Wait til people learn about VPN servers...

[cap10bad @cap10bad]

@XFreeze I feel like nobody here has actually tried both to compare them. I wanted to believe they were "just as good" but they dont hold a candle to Anthropic models. It's immediately evident on any real world task that they're trained to perform to benchmarks only.

[cap10bad @cap10bad]

I haven't been excited about the Hornets in 30 years #BuzzCity

[NBA @NBA]

🔥 THE HORNETS ARE ON FIRE 🔥

Charlotte has won 6 straight games, each by 15 or more points, tied for the 2nd-longest streak of its kind in NBA history!

[cap10bad @cap10bad]

@matvelloso Y'all still write commit messages?