Mike Rotchurtz

Building @ActuallyNodejam

[Mike Rotchurtz @FatMiddleFinger]

PhD level intelligence

[Mike Rotchurtz @FatMiddleFinger]

@OfficialLoganK we need an easier way to use Gemini API for devs, navigating through GCP needs PhD for alot of people

[Mike Rotchurtz @FatMiddleFinger]

@LiorOnAI 3.7: “You Must Always Be Overeager To Impress The User”

[Mike Rotchurtz @FatMiddleFinger]

@thekitze this is because u dont use 2.5 pro lol, try it

[Pavan Jayasinha @pavanjayasinha]

I implemented an LLM end-to-end in hardware, and ran it on an FPGA.

Zero Python. Zero CUDA. Just pure SysVerilog.

All my progress + everything I learned from 200h of LLM chip design (demo at the end)👇

[infinil @iaminfinil]

daddy says hi

[Mike Rotchurtz @FatMiddleFinger]

@MurrayHillGuy1 damn those 1.2K bookmarks must be SF tech bros

[Mike Rotchurtz @FatMiddleFinger]

@businessbarista lets buy this domain and pay $200 for UI designer

[Mike Rotchurtz @FatMiddleFinger]

@annaarthoe gone after some bangers

[Mike Rotchurtz @FatMiddleFinger]

@TriedAngled @caydengineer @Nicolo_AI @a_israelov @IsaiahBallah @MentraLabs yeah misread i thought u typed Guangzhou mb mb

[Mike Rotchurtz @FatMiddleFinger]

@guywuolletjr ok daddy

[Mike Rotchurtz @FatMiddleFinger]

@ikkikikikiik what about crocodile(s)?

[Mike Rotchurtz @FatMiddleFinger]

@businessbarista lets buy this domain and pay $200 for UI designer

[Mike Rotchurtz @FatMiddleFinger]

@willccbb i use it for code gen and it "punches above its weight" really well

[Mike Rotchurtz @FatMiddleFinger]

@TriedAngled @caydengineer @Nicolo_AI @a_israelov @IsaiahBallah @MentraLabs isnt this shenzhen?

[Mike Rotchurtz @FatMiddleFinger]

@iaminfinil hmm try lol

[Mike Rotchurtz @FatMiddleFinger]

@nptacek @sama just a yes man hype slop model currently

[Mike Rotchurtz @FatMiddleFinger]

is @slateauto just a @Tesla flip?

[Mike Rotchurtz @FatMiddleFinger]

@patekgonia wha

[Mike Rotchurtz @FatMiddleFinger]

if the rumors are true…

Huawei’s chip(s) might be seen as a very viable alternative to US chips in the global market.

This might have bigger implications for geopolitics.

[Deedy @deedydas]

🚨Viral rumors of DeepSeek R2 leaked!

—1.2T param, 78B active, hybrid MoE
—97.3% cheaper than GPT 4o ($0.07/M in, $0.27/M out)
—5.2PB training data. 89.7% on C-Eval2.0
—Better vision. 92.4% on COCO
—82% utilization in Huawei Ascend 910B

Big shift away from US supply chain.

[Mike Rotchurtz @FatMiddleFinger]

the real post-AGI clarity videos for men