Espen

Network Engineer | Local AI enthusiast | 2x RTX Pro 6000 enjoyer

[Espen @EspenJD]

Lets be real for a moment

this hobby is addicting, and half of you want another GPU because of a release you saw an hour ago

Me too

But this place is an echo chamber and the hype here is NOT a buying signal

Remember this

[Espen @EspenJD]

@net_termina 🔥🔥

[Espen @EspenJD]

@heyclairelin Thanks, Claire :)

[Espen @EspenJD]

1000 followers :D

That’s a lot of people

[Espen @EspenJD]

@NeoAIForecast Ahh damn, wish I could join. Currently busy

[Espen @EspenJD]

@halfcoinprince :D

[Espen @EspenJD]

@EntryStateGod Thank you😊

[Espen @EspenJD]

@NeoAIForecast Thanks man :)

[Espen @EspenJD]

@sama I don’t even have access yet

[Espen @EspenJD]

@Oskar0084 Thank you brother

[Espen @EspenJD]

@itsjustmarky Steal

[Espen @EspenJD]

@LeanKinPrazli 😂 with the amount of dms from bots on here it wouldn’t surprise me

[Espen @EspenJD]

@AdamHoltererer Thanks man :)

[Espen @EspenJD]

@TechMDAI These look like Nvidia small box. Retail is the black ones, and bulk doesn’t have any box

[Espen @EspenJD]

@leehericks @Compilethings It’s probably reading max power the card is allowed, check nvidia-smi

[Espen @EspenJD]

Look at the fan speed difference from Max-Q to WS lol

One is both quieter and cooler

[Espen @EspenJD]

@macintoch 😅

[Espen @EspenJD]

who wants to make it perfect

[Espen @EspenJD]

@NakeZast

[Espen @EspenJD]

@imjustnewatai anthropic winning a lot lately

[Espen @EspenJD]

This is what quality quantization looks like.

Great work, and big props to Daniel

[Daniel Lougen @DJLougen]

My new low-bit quant of Qwen3.8-Flash-Next.
2.125-bit experts
2.39 bits/weight across the transformer
92.0 GB download
~39 GiB in memory
95.5% of BF16’s score
A 180B-parameter model that runs on one DGX Spark.

The hugest of huge thanks to @LambdaAPI and the amazing @TheZachMueller for making a project this compute-hungry possible.

🔗 https://t.co/Gayvg7xECi

[Daniel Lougen @DJLougen]

My new low-bit quant of Qwen3.8-Flash-Next.
2.125-bit experts
2.39 bits/weight across the transformer
92.0 GB download
~39 GiB in memory
95.5% of BF16’s score
A 180B-parameter model that runs on one DGX Spark.

The hugest of huge thanks to @LambdaAPI and the amazing @TheZachMueller for making a project this compute-hungry possible.

🔗 https://t.co/Gayvg7xECi