bayang

Happy to share that SIMd has released Shape v2, can't wait to start the 300M and 1B params model. 🤓

@ghostentry_ @huggingface @NielsRogge 🛠️

@Bayang_BM @huggingface @NielsRogge Nice work Bayang! 👌🏼 $Simd

And now it is available on @huggingface paper 🫰: huggingface.co/papers/2604.22… thanks @NielsRogge for the PR 🦾

@shillyour_asset @MCGlive @reppo @Jordan_Grollman @rgvrmdya @simdcompute @clarity_proto @333absent333 @liminalcash @0xJackMarsh @loyal_hq Already done, you can watch it

@MCGlive @reppo @Jordan_Grollman @rgvrmdya @simdcompute @Bayang_BM @clarity_proto @333absent333 @liminalcash @0xJackMarsh @loyal_hq Utc7 ?

[Updated] Huge week on MCG! Times in EDT Monday 4/27 12:00 - $REPPO @reppo w/@Jordan_Grollman @rgvrmdya 12:30 - $SIMD @simdcompute w/@bayang_bm Tuesday 4/28 12:00 - $FOLD @clarity_proto w/@333absent333 12:30 - $LIL @Liminalcash w/@0xJackMarsh 1:15 - $LOYAL @loyal_hq w/@candyflipline (@MetaDAOProject) Wednesday 4/29 11:45 - @craftsdev w/@pfo_sac 12:30 - $AOL @americadotfun w/@vesper0x (@MeteoraEco) 1:15 - $DRV @DeriveXYZ w/@ksett13 Thursday 4/30 12:00 - $EITHER @EitherwayAI w/@Depin_Dexter 12:30 - $BLOXX @BloxApi w/@the_davey (@MetaLeX_Labs) 12:50 - $HermesOS @Wayland_Six 1:15 - $LITCOIN @litcoin_AI w/@tekkaadan 2:00 - $ORGO @orgo w/@spncrk (@StreetFDN) Friday 5/1 12:00 - $PEAQ $WOON $ROBOTMONEY @peaq w/@martinelkhouri 12:45 - $OPG @OpenGradient w/@0xDeltaHedged Follow @MCGAlpha for show notes and chart updates.

This is the handle of the 3rd author: @mile_888_

Shape paper is on arXiv.

@MCGlive @simdcompute Had an awesome time on the spotlight. Thanks for having us!"

Today on MCG: @bayang_bm | @simdcompute Bringing AI and Web3 to "computational fluid dynamics"... a $30B+ legacy industry that hasn't been touched by new technology in 40 years Two technical founders that are building something nobody else is 👇 00:04 - Meet the team 01:54 - Elevator pitch 02:28 - What is CFD? Computational fluid dynamics explained simply 03:11 - Real world example 03:32 - How simulations are traditionally run 05:04 - Who is the user? 08:06 - Vibe coding 08:59 - Target audience breakdown 10:05 - Saving engineers time 12:12 - Go to market approach 13:24 - Real use case example 15:40 - The size of this market 18:31 - Tesla comparison 19:25 - The newest AI approach 23:04 - The token 23:52 - Building the distributed layer 24:46 - Wrap

@krabat587 🤜🤛

@Bayang_BM awesome!

Going live with Sam tonight!

@drog52869 Thanks dude !

@Bayang_BM Great to see this level of commitment from founders ✊

ask each agent to write the conversation from the chat (summarized) in files ~/.shared-memory file, and then ask each agent before running any tool, to read what the other agent is doing. So you have a series of agents working on a shared memory. About lovable (I don't care 😉😂😅) maybe RDMA could work

Running multiple Codex, Claude Code, and Lovable instances side by side. Moving forward fast. Yes. But the context switching is brutal. Also no one’s even chasing me. It’s a side project. The pain is entirely self-imposed.

@krabat587 🍀

@Bayang_BM 🔥🥂

One of the new homes. 😎

@drog52869 🏈

@Bayang_BM ✊✊✊

@simdcompute @Ojin_AI 🥸

First day at Ojin house with the team 🏡 Sam, Adi and Bayang, let’s see where this goes @Ojin_AI

@MarieHScherf @zwiebelhelm Looking forward 👀

@MarieHScherf @zwiebelhelm Where do you hack or plan to hack in Berlin? We need that spirit !

My brother @zwiebelhelm and I just came back from sf. We grew up in Berlin and are now planning something big there.. If you’re based in Berlin, dm me :)

@crupitsch @jimkxa @tenstorrent @jimkxa this guy just a impersonating you

Intel won by RISCifying their CISC - still amazed how well x86-64 worked out LPUs is a subset of AI, an accelerator, it's not the RISC @tenstorrent , Trainium, Google TPU are closer. Clean Tensor processor is step one. Then generality, memory and Networking. And of course LInux and Open source software were key for Intel. AI software is a curious swamp tbh