Calbytez

solo founder · 4 products
🏆 SLUGGER · 61.2% MLB HR picks · 8pm ET daily
🍳 PANTRE · 🩺 MEDLEARN · 🤖 Norevel
DMs open

[Calbytez @calbytez]

1 founder · shipping 4 products on Claude Code:

🏆 SLUGGER — MLB HR picks, 61% all-time
🍳 PANTRE — AI cooking
🩺 MEDLEARN — med-ed (in flight)
🤖 Norevel — AI-native ops for SMBs

Daily build logs · Daily picks · Receipts.

ET · solo · DMs open

[Sudo su @sudoingX]

from a 12gb gaming card to 2x dgx spark, here's the model i'd run at each tier, the speed i measured and the settings that make it work, bookmark this:

> 12gb: bonsai 2 27b ptq1_0 with the mtp head, 50 tok/s fresh on a 3060. kv cache at q4_0, one slot, n-max 1 on the head and reasoning effort medium, because the default xhigh can think through your whole token budget and hand you an empty answer.

> 24gb: qwen 3.8 27b q4_k_m, 41.1 tok/s on a 3090 with mtp off and 65.8 tok/s with it on, at a 131k window. add --spec-type draft-mtp --spec-draft-n-max 2 --parallel 1 to llama-server and keep kv at q4_0, f16 kv hits a wall between 90k and 131k on 24gb.

> 1x dgx spark: moes, since it has 273 gb/s of bandwidth. qwen 3.8 flash next ud-iq4_xs does 44.4 tok/s in llama.cpp and ling 3.0 flash on the official int4 does 38.7 tok/s in vllm, both with mtp on.

> 2x dgx spark over connectx: qwen 3.8 flash next official fp8, tensor parallel across both boxes, 45 tok/s fresh and 35 tok/s mid-build with mtp, full 262k context with vision.

the rule under every tier is the same, decode tops out at your memory bandwidth divided by the bytes read per token, which is why a 1.75 bit file flies on a 3060 and why a 180b moe that only reads its active experts per token flies on a spark.

[xEP Network | Sports Analytics @xEP_Network]

Pete Crow-Armstrong 1+ HR at +320 carries a 0.73 projection 💣

+28.2% xVALUE and an 80 HR Power. He grades 92 on the sinker and 84 on the cutter.

Bryce Miller throws those 13% combined. The four-seam at 41% grades 69.

Miller sits at a 3.71 ERA with a .224 allowed, and the overall matchup is a 63. Big price on a bat that has to find the pitches he sees least.

[Calbytez @calbytez]

@NVIDIAGeForce #RTXPowersPlay

[kira 👾 @kirawontmiss]

this might be my favorite tweet this year😭

[Calbytez @calbytez]

My laptop died at midnight. Not one agent noticed.

They don't run on my machine. They run on a server that never sleeps. The laptop is just a window into the company.

That's the difference between using AI and running your business on it.

[Calbytez @calbytez]

@Motionfly_co Need it

[Calbytez @calbytez]

@georgesttock Want

[Calbytez @calbytez]

@sweepsfinalboss CASE STUDY

[Calbytez @calbytez]

@blvckledge Swipe

[Calbytez @calbytez]

@niyetsel Claimed

[Maestro🍀 @MaestroHR]

I AM BUYING HENNESSEY TODAY

[Calbytez @calbytez]

@Zephyr_hg NEWA

[Calbytez @calbytez]

@mynameisyahia @getcontextdev Scraping websites and company enrichments. , I feel like context dev may be the one use api I can start using over apify and others if information is rich [email protected]

[Calbytez @calbytez]

@mynameisyahia @getcontextdev Heyooooo let’s get to crawling

[Calbytez @calbytez]

@johnvirality Augc

[Calbytez @calbytez]

@Berry_cooool Auto

[Sammy Smacks @SammySmacks26]

It’s their time of the year, mine too. 🏄🏽‍♂️

[Calbytez @calbytez]

@itsalexvacca Repo

[Calbytez @calbytez]

@minchoi Mine hasn’t been working

[Calbytez @calbytez]

@thenowhereway Hey man Q