LocalAIMaxxer // Rezi

Founder @ Speaker
Local AI Enthusiast - M3 Max 128gb, 2x RTX 5080
Open to AI Projects / Work
C#, C++, Unity, Computer Science 1st Class

[LocalAIMaxxer // Rezi @RZY333]

Moderating a panel on 22nd September in Hong Kong on AI in Financial Markets with -

-Generative AI Ambassador & APAC Lead, HSBC
-Head of Research and Development, BANDS Financial Limited
-Senior Strategic Lead, Financial Services, Alpha Sense

Potential spot for a 4th, dm me.

[LocalAIMaxxer // Rezi @RZY333]

@alexocheema are we sure that qwen3.6 is the most intelligent model that can run on a M4 Max? ..

[LocalAIMaxxer // Rezi @RZY333]

Made with GPT Astra? :P

[OpenAI @OpenAI]

Since announcing Jalapeño, our first custom inference chip, we’ve been testing it and the system around it.

The results show a major advance: more intelligence from every watt and faster responses, delivering both higher throughput and lower latency in one architecture without sacrificing efficiency.

[LocalAIMaxxer // Rezi @RZY333]

@Apple If you send me an M5 Ultra I promise to buy 100s more in the future. I'll personally make sure to increase apple sales and non stop glaze apple products to everyone! Consider this part of your marketing budget 😀 pls. no really. @tim_cook

[Ivan Fioravanti ᯅ @ivanfioravanti]

I want to thank @Apple for the M5 Ultra 512Gb that they’ll send me!
(Trying the self-fulfilling prophecy way 😉)

[LocalAIMaxxer // Rezi @RZY333]

@Youssofal_ What could I cook with 1x 5080 and 1x 3080 ?

[LocalAIMaxxer // Rezi @RZY333]

@alexocheema Yeah will also need a second mortgage to buy 4 m5 Ultra 512gb

[LocalAIMaxxer // Rezi @RZY333]

@jun_song @Youssofal_ how do we get mtplx of this ?

[LocalAIMaxxer // Rezi @RZY333]

@Tono_Ken3 Mlx?

[LocalAIMaxxer // Rezi @RZY333]

@support_huihui Mlx?

[LocalAIMaxxer // Rezi @RZY333]

@MiaAI_lab What about 5080 users are we forgotten 😢

[LocalAIMaxxer // Rezi @RZY333]

@reczko_konrad @iwoplaza can we add this in the Mac camera to improve lighting

[LocalAIMaxxer // Rezi @RZY333]

@elliotarledge honestly I don't know yet, I guess if you are using terms that aren't standard english, but technical terms related to coding etc. it will clean it up wrong but that's expected with a small model. The speed just feels so nice and snappy so idc.

[LocalAIMaxxer // Rezi @RZY333]

186 Tok/s Peak , 80 Tok/s Average, 130k Context, Ornith-1.5-35b-a3b in Pi on M3 Max Macbook, please tell me Qwen3.8-35B-A3B will be this snappy.

[LocalAIMaxxer // Rezi @RZY333]

@trevorwood222 I didn't play with it too much, can you put any model in it to test or?

[LocalAIMaxxer // Rezi @RZY333]

@superalesha Don’t worry I’m feeling like that about my 128gb m3 max MacBook that I spent $4.5k on 😭

[LocalAIMaxxer // Rezi @RZY333]

@ashxhart Interested, I'm dying to find a local model to use in Pi that actually doesn't break on real agentic coding work.

[LocalAIMaxxer // Rezi @RZY333]

@ddalcu As someone with a 128gb M3 Max Macbook and a RTX 5080 PC, it would be great to have something to allow me to join the two.

[LocalAIMaxxer // Rezi @RZY333]

@MarshallJGould Must be the gtx 5090, most of us with rtx 5090 can’t get those speeds 🙈 we believe you cuh