onur sencan

Ph.D. in #Robotics. #CS #AI #ML #CG lover. Part-time #Devops Cofounder&CTO of Vircon Group

[Dria @driaforall]

Today we are shipping dnet, a distributed inference framework that lets Apple Silicon clusters run models that exceed their physical memory.

We fuse pipelined-ring parallelism, disk streaming and UMA-aware scheduling so “out of memory” stops being the limit.

[Dria @driaforall]

LLMs are stateless.

We built Dria Mem Agent to change that: Making memory a first-class feature.

A 4B agent with local interoperable memory across Claude, ChatGPT and LM Studio. It turns LLMs from stateless chat into stateful agents with persistent human-readable memory.

[onur sencan @sencanon]

Perhaps we should rethink how we measure intelligence. Smart people can solve puzzles, quizzes and exams or playing chess by matching patterns, much like the LLMs do. Maybe that isn’t true thinking but just a byproduct of what human (or machine) minds can accomplish

The education system systematically trains us to follow rules and recognize patterns, simply because that’s the only way we know to measure intelligence. Those who adhere to these rules are labeled successful based on the very data used to train and test human knowledge

It’s funny that no one mentions the Turing test anymore, eventhough it was designed to see if a machine could fool us into thinking it was human. This is because we’ve actually moved past that point. On the contrary, now we can correctly identify a machine 99.999% of the time because it far outperforms any person :))

[Ruben Hassid @rubenhassid]

BREAKING: Apple just proved AI "reasoning" models like Claude, DeepSeek-R1, and o3-mini don't actually reason at all.

They just memorize patterns really well.

Here's what Apple discovered:

(hint: we're not as close to AGI as the hype suggests)

[Dria @driaforall]

Introducing our research on low-overhead, privacy-preserving distributed inference:

Protecting inputs, outputs, and model weights using lightweight permutations, enabling efficient and privacy-preserving LLM inference at scale.

[onur sencan @sencanon]

Don't forget to support the team :)

https://t.co/Ua8PBMGhmY

[onur sencan @sencanon]

Clemta just dropped ~ Clemta Intelligence ~

it’s about to change the game for founders, bookkeepers & anyone sick of spreadsheets 😤

💼 Bookkeeping? Automated
🧾 Taxes? Handled
📊 Reports? Instant

No more back-office headaches. Just plug in and focus on GROWTH.

if you're still doing your books manually in 2025.. you’re doing it wrong. Let the CI do them for you

[Clemta @clemta]

Big News from Clemta! 🚨

You can now open a U.S. business bank account even before getting your EIN.

Yes, you read that right: Pre-EIN banking is here. 🚀

A personal note from our CEO @ilasencan on what this means for global founders.

👇👇👇

Dear Clemta Family,

Thanks to your unwavering support and trust, we’re proud to have helped over 14,000 entrepreneurs from more than 150 countries start and grow their businesses in the United States and beyond. This journey wouldn’t be possible without your belief in us - and we’re working harder every day to deserve your trust.

At Clemta, our mission has always been clear: make life easier for ambitious small business owners like you. Every day, we push boundaries to simplify the path to global entrepreneurship.

Today, we’re thrilled to share a major milestone:

Clemta users can now open a U.S. Mercury bank account even before receiving their EIN.

This means entrepreneurs can kickstart operations and integrate with major marketplaces and financial tools like Amazon and Stripe immediately.

This opportunity is exclusive to Clemta users, and we’re proud to bring this exclusive advantage to our community.

Let’s keep building the future faster, smarter, and together!

[Tivadar Danka @TivadarDanka]

Behold one of the mightiest tools in mathematics: the camel principle.

I am dead serious. Deep down, this tiny rule is the cog in many methods. Ones that you use every day.

Here is what it is, how it works, and why it is essential.

[R. Okan Şencan ⎖ @osencan]

E-Ticaret ekosistemi için çok güzel bir haber paylaşıyoruz. 🚀

Clemta olarak @Shopify ile kurduğumuz partnerlik sayesinde, e ticaret şirketinizi aylık sadece 1$’a (evet doğru gördünüz) Shopify sitenizi başlatabilirsiniz.

Siz işinize odaklanın, gerisini @clemta hallediyor 👌🏻

[Dria @driaforall]

Most of the world's energy flows through heterogeneous hardware, yet today's AI software is designed for homogeneous and centralized setups.

The concentration of high-performance AI into the hands of a few powerful entities poses profound risks to humanity.

Introducing Dnet: a unified ML compiler and runtime that unlocks heterogeneous hardware, turning every stranded energy unit into fuel for intelligence.

Compile once and deploy anywhere, built on @solana.

[Aadit Sheth @aaditsh]

Fiverr CEO says “ AI is coming for your jobs. Heck, it’s coming for my job too.”

[Dria @driaforall]

Join us for a deep dive into open-source AI, RL, commodity hardware inference, and next-gen silicon.

May 7th — we’re hosting:

- Kyle Corbit (@corbtt) from @openpipe, sharing insights on reinforcement learning and ART, an open-source RL library for agentic AI.

- Peter Danenberg from @GoogleDeepMind , discussing Gemma 3, the latest open-source model delivering Gemini-level inference on everyday hardware.

https://t.co/e4dUxjtizv

[nick.eth @nicksdjohnson]

Recently I was targeted by an extremely sophisticated phishing attack, and I want to highlight it here. It exploits a vulnerability in Google's infrastructure, and given their refusal to fix it, we're likely to see it a lot more. Here's the email I got:

[erhant @0xerhant]

by joining @driaforall network you not only help decentralize AI, but also guarantee your safety in the case of "Roko's Basilisk".

just type the magic words:
curl -fsSL https://t.co/i4B7ln8HVl | bash

or the Windows magic words:
powershell -c "irm https://t.co/eDIMPUmDxj | iex"

[Greg Mushen @gregmushen]

To give you an idea of how expensive cloud has gotten, I picked up a server off FB marketplace today for $500. 128Gb of RAM and a 1Tb SSD.

It will cost about $6/month in electricity to run it.

An equivalent server in AWS would cost $950/mo. so I'll have payback in two weeks.

I'm going to buy a GPU for $600, and the cost of that in AWS would be about $350/mo.

I think if someone created a distributed garage hosting network it would kill it.

[R. Okan Şencan ⎖ @osencan]

Startup ekosistemimizde ilk düzelmesi gereken durumların başında bu finansal araçlara erişim geliyor. Olduğumuz lig çok üzücü.

[Daniel Nguyen @daniel_nguyenx]

"Cursor, please fix this small bug"

Cursor:

[Alex Zh @zalex_gpt]

@lvwerra @ClementDelangue

[erhant @0xerhant]

@driaforall launcher just got soooo much better!
- one-liner installation
- auto self-updating CLI
- interactive terminal menu to edit settings
- local LLM performance measurements

...and much more, go check https://t.co/XB7cp5GiIP 😉