Mass

Why dual RTX 3090 is a perfect starter pack into local AI - 48GB VRAM for ~$2000, just drop them into your pc - Runs Qwen3.6-27B at Q6 with fat context - CUDA + vLLM/llama.cpp work out of the box, zero headache - Buy threadripper and scale your rig up to 8x3090 - Prices are structurally rising, supply demand shock 3090 is the people's card

@samirettali Your life will be divided into before and after

time to try out hermes agent

@samirettali What GPU are you running?

i've been running qwen 3.6 27b locally for a couple days and man it's so much fun it made me want to build a dedicated inference machine, it's going to be a good weekend of research

@Snixtp Opus also has the same logic with failed tests, looks like a widespread issue

I think memory in Codex is genuinely one of its best features. Being able to reference past chats, previous decisions, and project context makes it feel way more consistent over time instead of starting from zero every session. But there's also a downside I've noticed: If you test something once and it fails (even if it was just a one off issue, bad timing, wrong environment, or temporary bug) Codex tends to permanently internalize that result. After that, it often refuses to even reconsider the approach and just says: "This didn't work before, so it's probably not worth trying again." I'm on the fence turning memory off for a while to see if Codex opens more doors/is more open minded on issues

@chooserich It’s amazing that this meme stays relevant so long lol

Ya’ll sold?!?!

@witcheer Gl gl

the full spec for my headless Linux server: - GPU: MSI RTX 5090 Gaming Trio OC 32GB - CPU: AMD Ryzen 5 9600 (Zen 5, 65W) - Mobo: Gigabyte X870 Gaming X WiFi7 - RAM: 64GB DDR5-5600 CL36 - Storage: 2TB NVMe - PSU: 1200W 80+ Gold - Cooler: Arctic Liquid Freezer III Pro 360 - Case: SilverStone SETA H2 - OS: Ubuntu Server 24.04 LTS and I SSH the whole thing to my Mac. my goal is building an inference and fine-tuning machine that runs 24/7.

@NeoAIForecast True

@MemoryReboot_ Qwen is the Goat!

Qwen3.5 27B released 24.02.2026 Qwen 3.6 27B released 22.04.2026 Qwen 3.7 27B released .....

@stevibe Time flies 😭

Wait, the first 27B was released only 90 days ago?

Why not try this with iPads next time

@chirag @sudoingX Delete this

@MemoryReboot_ @sudoingX this doesn't look to be open weights though. :|

qwen is unreal. they just dropped 3.7 max and it is beating opus 4.6 max on most of the benchmarks they ran. terminal bench, mcp use, math, instruction following, humanity's last exam. and the apex math number, 44.5 against opus 34.5, that is not a small gap. the 35 hours straight on a kernel optimization task with 1000+ tool calls is the part i keep rereading. that is the agent era thing actually happening, not a slide. the speed alibaba is shipping at right now is the whole story, 3.6 was last month, 3.7 max today, nobody else is moving like this. one thing though, please open source this one too. 3.6 dense made the entire local llm ecosystem better. the max tier going api only would close a door we have been keeping open. give us the weights eventually.

@jun_song Crypto is great, 90% people hating on it are just salty they couldn't make any money there

To my 12,000 followers, I completely understand if you are confused or disappointed by my sudden posts about crypto. Those who already have strong negative feelings probably won't accept anything I say right now. However, I have spent the last few months deeply studying the open source ecosystem, meeting countless developers, and listening to their struggles. The reality is that the open source funding system is completely broken. The top 0.1% of high-profile creators hoard the vast majority of sponsorships, regardless of how critical their project actually is. The rest of the developers are barely getting by, even though many of them have made breakthroughs that could literally change the world. I was incredibly frustrated by this broken system, and I kept searching for a way to fix it. I came to the conclusion that crypto mechanics are actually the most effective tool for this. So, I decided to risk my entire reputation and face all the stigma head-on to flip this into a positive ecosystem. Just like I previously worked to change the narrative around local LLMs and Sovereign AI when they were dismissed. Open source must win, and we don't have much time. If you are too disappointed, you are welcome to unfollow me. But I will welcome you back later, after I have radically revolutionized this deeply rooted problem. I am sorry to have let you down. Sincerely, Jun Song

I wonder how many people roped over the emergence of AI

@elielAGI 😭

@MemoryReboot_ should probably have thought about the acronym for just a bit more

@elielAGI It's ASI

@MemoryReboot_ at this rate qwen 4.0 27b is super agi

@witcheer Congrats, 5090 is a beast 🔥

I’m proud to announce that I bought a GPU hopefully the start of a wonderful journey. I ran a lot of benchmarks on my RTX 4060 8GB, which helped me learn the basics and understand how it works under the hood. now we scale up!

@micheltamanda Facts!

@MemoryReboot_ The entire local AI community is waiting.

@misbah_bangee Why not

@MemoryReboot_ 20.06.2026?

@DrTBehrens 🤝

@MemoryReboot_ 22.6. I am ready.

@TerrorCosmic Manifesting

@MemoryReboot_ AGI by the 8th of August

@chooserich Choose rich not cancer 👊

If the number of people offering prayers is any indicator… I’m absolutely going to destroy this cancer 😭💪

@aadityaa_26 @sudoingX I hope

@MemoryReboot_ @sudoingX is there any news about that?? will they drop oss??