Sufyan

@VraserX I don’t think open source will truly ever be better or catch up to closed source simply because of how much they are pouring in, however that doesn’t mean open source can’t innovate and create new and stronger ways of serving ai

I keep wondering if open source AI can really catch up long term. The compute gap is already massive, and closed labs are pouring insane money into training. Right now, open models can cheaply copy a lot through distillation. But what happens when closed labs finally figure out how to block that? At some point, open source may stop chasing the frontier and start chasing shadows.

@matthewmillerai Local will win.

Fable 5 was only the beginning. Now GPT 5.6 AND Gemini 3.5 Pro are delayed. I care about vibe coding and AI a lot. I am so frustrated.

@bridgemindai Local is the way.

The U.S. government is destroying AI right now. Fable 5 has been completely suspended (no updates). GPT 5.6 has been delayed (twice now). AND Gemini 3.5 Pro has been pushed back. This isn't just about Fable 5 anymore. It's about all frontier models.

@beffjezos Starmind is a very cool name.

StarMind is a banger. Whoever came up with that at xAI needs a raise...😉 🚀🌌🧠🫡

Once Starmind goes live, all the attention will shift from land-based data centres to Starmind. Calling it now.

@CardilloSamuel I would love to serve local inference to regulated industries

just had a call with yet another company switching to local ai. they were using Qwen 3.6 35b served through my spark so they could test their pipelines through it and see if its a fit and what i heard in that call is pretty crazy and i feel it is something that most anti-local ai misunderstand. what the customer told me is literally that they kept comparing the results between Claude, ChatGPT and the locally served Qwen 3.6 and largely preferred the results they were getting via Qwen - insane! the plan is now to deploy a DGX Spark in their office for the MVP + a 6000 pro blackwell-enabled machine for training & later on they're looking to basically buy a DGX Spark per customer they have so they can serve each customers an optimized finetune. this is the future. literally. and once again, this is the worst it'll ever be.

@lalitgrateful @pranavcodes_ Don't drink 6 espressos a day and train 3 hours a morning on no carbs and bad sleep.

@pranavcodes_ So where did you fuck around and What did you find out?

Let's lock in 🔐

How to make a UK heatwave even more unbearable 🥵

@man_AV_ But then you can argue that experiencing those things are important for human development and you wouldn't be who you are without them.

People are more comfortable talking to an AI than humans because:- 1. It won't judge you 2. It is not biased 3. It will not gossip 4. It will not use the info against you in an argument.

@kapilansh_twt Sakani AI

why is Japan so behind in AI they gave us robots bullet trains PlayStation and yet barely show up in the AI race what actually happened ?

@IceSolst Only problem with it is that either it is super slow or unusable

You can bypass ISP restrictions: 1. Download Brave browser 2. Start a new window with Tor (built-in support with Brave)

@sudoingX Nothing Im buying a spark pronto

what's stopping you from buying a gpu dude?

I hate when people say 'if you haven't got anything to hide, why do you care'. My friend it was never about hiding anything. what if you simply don't want to be looking over your shoulder 24/7.

I'm going through this weird transition of vibe coding addiction to just utter vram addiction, like why do i suddenly want to hoard mountains and mountains of vram. Blows me away.

@josepha_mayo As soon as you have the 100gb ur going to want more and more.

100gb of vram would be heaven to me

I think Rust will continue to keep climbing in terms of popularity in this age of AI.

@icanvardar To be honest I don't see a reason why Russia dont have frontier ai.

it’s kinda wild that US and China are shipping frontier AI models while the rest of the world is AFK

@jun_song 2 bit must be even better!

1-bit quantized GLM-5.2(lowest quant quality) has the best quality among frontier models. Really impressive.

@jvr0x @MiaAI_lab @grok Sure you can! Go to's are hermes agent and grok build

@MiaAI_lab @grok Can you use grok build with local models? 👀

That’s great news. /goal is now in @Grok build. I use Grok with my local models pretty often, gonna test it out.

@MiaAI_lab Deserved!

Just realized I crossed 2k followers. Thank you so so much 🥹