Lalo

building. writing. // prev @microsoft @umich

[Lalo @Laloquidity]

https://t.co/tx1zqT1KHE

[Lalo @Laloquidity]

The US Government & Anthropic inadvertently accelerated open source AI adoption and development by multiples through its suspension and zero communication on Fable's restoration.

[Lalo @Laloquidity]

@majinsayan Old friend

What are good resources about Sufism?

[Lalo @Laloquidity]

@AlexFinn what was the main argument from the gpu folks?

[Lalo @Laloquidity]

@Zakariyalyqz we need more of this, finance built on true economic value and not credit creation

[Lalo @Laloquidity]

@amasad @Replit

[Lalo @Laloquidity]

@SafdarAlam Found this gem on your youtube channel. What a great breakdown 😂 https://t.co/JEhnDmagRg

[Lalo @Laloquidity]

Noted on terminology. The stack you're outlining is where I've spent the last 5-6 years. Built a Layer 1 blockchain with governance, a consumer neobank with card and banking rails, FX, bridges, on/off ramps, crypto native institutional desk relationships.. Happy to merge efforts if it ever makes sense. Will read your upcoming works.

[Lalo @Laloquidity]

@SafdarAlam That is end game! Productive enterprise, musharakah.. Excited to read the upcoming work. The framework deserves proper onchain infrastructure. If that's what you're building toward, I can help. Working on pieces of this already.

[Lalo @Laloquidity]

Agreed. Existing models are workarounds within the conventional system that replicates economics of debt. Onchain is the first real chance to build from first principles rather than port the same debt architecture. Previous halal stablecoin attempts lacked use cases and distribution that would net institutional flows. It should be part of a full ecosystem. I want to build It right. Following your works closely.

[Lalo @Laloquidity]

https://t.co/gzYOIx1cyT

[Lalo @Laloquidity]

@SafdarAlam Was the supposed 125% collateral fake? Almost like blockchain could help solve that..

[jack @jack]

https://t.co/jgZkBvYOPt

[Lalo @Laloquidity]

Work at the speed of thought.

Fully local speech-to-text:

67ms for a sentence on an M3 MacBook → Pastes directly into your active app → Filters out uh, um, filler words → No cloud, no API, no subscription.

Link: https://t.co/tMLtFQhJZT

[Lalo @Laloquidity]

https://t.co/ZSGwW3r8Xh

[Lalo @Laloquidity]

"it's so easy a Qwen could do it"

prob time to touch grass if you understand that

3.5 27b is solid though esp w/ hermes

heterogeneous hardware p cool step forward

[the tiny corp @__tinygrad__]

If you have a Thunderbolt or USB4 eGPU and a Mac, today is the day you've been waiting for! Apple finally approved our driver for both AMD and NVIDIA. It's so easy to install now a Qwen could do it, then it can run that Qwen...

[Lalo @Laloquidity]

@LottoLabs @daniel_mac8 perfect thank you !

[Lalo @Laloquidity]

@LottoLabs @daniel_mac8 @LottoLabs do you possibly have a writeup on your optimal setup w/ 3090? after reading your posts thinking llama.cpp + qwen3.5 27 + hermes