Tjep

AI and stuff

[Tjep @BBXIntelligence]

@KhalidAlNumi 2x nvidia dgx station

[Tjep @BBXIntelligence]

@AidenAllred2 @LongFormMath

[Tjep @BBXIntelligence]

Keep it local. Your work and ideas are not safe

[Mehdi (e/λ) @BetterCallMedhi]

go fuck yourself @sama

claiming that you solved navier stokes because 10000 agents ran in circles for 88hours on a multimillion dollar gpu cluster to formalize in lean a blowup case under controlled external forcing is pure scientific vulgarity the clay mathematics institute millennium prize does not ask m whether you can artificially force a singularity in a fluid by injecting an ad hoc smooth external forcing term f(x,t) to twist the vortex until it breaks the real problem questions the fundamental stability and global smooth existence for 3dimensional incompressible euler & navier stokes equations under natural conservation laws and viscous dissipation alone using a mathematical loophole on forced equations to parade a century old victory is a major conceptual scam Altman

technically & epistemologically what you present as an agi breakthrough is nothing more than bruteforce combinatorial autoformalization the ai did not understand fluid mechanics it simply navigated a continuous search space previously mapped out and constrained by the monumental work of human mathematicians like tristan buckmaster/ levent alpöge / diego córdoba or tarek elgindi coordinating 10000 agents to check the logical consistency of a 100 page proof via lean is a software engineering feat and computational parallelization triumph not an intrinsic scientific discovery it is the victory of the compute bulldozer over abstract human intuition repackaged for the public as a higher mathematical consciousness

to this theoretical imposture you add a disgusting ethical and industrial cynicism taking advantage of private codex sessions and informal preprints from academic researchers to siphon their research leads and then trying to redact or erase the contribution of levent alpöge under the pretext that he works at rival anthropic is intellectual serfdom openai behaves like a feudal lord of silicon appropriating the cognitive subsistence of independent scholars threatening their careers behind closed doors if they protest and turning community academic labor into a privatized pressrelease

this entire staged event serves a desperate financial agenda in a pre ipo panic facing the slowdown of scaling laws and growing investor skepticism over the profitability of foundational models openai needs to manufacture an artificial sputnik moment claiming to solve a millennium prize without immediately submitting the proof to traditional peer review means using the prestige of fundamental mathematics as cheap marketing fuel to inflate a delusional valuation!!!

real science is not a clout chase on social media or a compute spike spent to rob the clay mathematics institute it is a quest for elegance physical truth and universal rigor to decode reality true artificial intelligence will not emerge from hostile corporate takeover of academic work hidden behind computational bruteforce but from architectures capable of generating new conceptual paradigms by masquerading constrained formalization as the collapse of physics greatest mysteries you did not solve navier stokes you only proved how far silicon valley will go to prostitute scientific integrity for capitalist spectacle

[Tjep @BBXIntelligence]

@sudoingX I recalibrated my garmin to sync to my own machine and stop sending data to Garmin.

[Tjep @BBXIntelligence]

@Blackfrost_AI Hit me up, and I’ll test it later

[Tjep @BBXIntelligence]

@sudoingX 3 dgx spark
1 5090
2 3090
2 3060

[Tjep @BBXIntelligence]

@Hikari_07_jp You should get a SPD TYPE 3 wall socket.

[Tjep @BBXIntelligence]

@dsjfnsdlkjfbsdl @TheAhmadOsman If you have that problem you should run fielddiag. If it fails you should be eligible for RMA.

https://t.co/sgMqOnz7Rh

[Tjep @BBXIntelligence]

@Hikari_07_jp I wouldn’t mind cheaper sparks.

It is still a amazing piece of hardware.

And ppl seem to forget that today is the worst your hardware will ever be at running AI.

[Tjep @BBXIntelligence]

What a day for open source.

[Tjep @BBXIntelligence]

Sooo.. i added another spark to the setup today ! Very excited to build with this setup ! @NVIDIAAI can you help me achieve that TP=4 ? 🥳 kind regards a Shadow lurking AI enthusiast

[Tjep @BBXIntelligence]

@Hikari_07_jp Or turn of heating in the house/apt and use it for heating the whole thing?

[Tjep @BBXIntelligence]

@MiaAI_lab any1 else with these?
-mhc_post_tilelang CUDA_
-ERROR_ILLEGAL_ADDRESS
-NVIDIA Xid 31
-MMU write fault

[Tjep @BBXIntelligence]

@BrianRoemmele @grok @bot Brian, a question for you.

It seems like you are all for local serving and keeping your data local.

How does this comply with that ?

[Tjep @BBXIntelligence]

@BoldingBuilds @Hikari_07_jp Another question ^^

Do you think it is possible to add 2 MBs on this rig?

[Tjep @BBXIntelligence]

@MiaAI_lab The future ☺️

[Tjep @BBXIntelligence]

@sfxnz @Hikari_07_jp Sometimes nightmare EVEN with cloud

[Tjep @BBXIntelligence]

Thank you qwen!

[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

[Joseph Sauvage @JoesInvestments]

Spent the last 12 hours somewhere nobody has been: GLM-5.2 on vLLM 0.27, four DGX Sparks, the brand new native sparse attention path for this hardware. Got further than anyone and hit a wall I can't move alone, so I'm handing the whole expedition to the community.

What I found: the native path is real. GLM is in mainline now, the new SM120 backend engages, and I got first tokens in 1.3 seconds with zero community patches. The entire overlay-and-patch era we've all been living in is about to be obsolete.

What's in the way: three bugs. I fixed two of them (both filed upstream with verified workarounds, vllm 51920 among them) and mapped the third precisely: on 4-node clusters the engine permanently stalls after a minute of idleness. Two-node setups don't hit it. Full reproduction filed as vllm 51921.

I walked so someone else can run. Everything is in my repo under v027: the fix layers, the launcher, the probe battery, and exactly where the frontier stops. It's on the TP=4 GLM community to pick up where I left off now. First one past that wall takes my mantle, and your name goes in the README.

https://t.co/MqhtA9tCBC

[Tjep @BBXIntelligence]

@Hikari_07_jp Its crazy. And imagine. This is still the early stage. 😅😅