Didier FRANC

Hacking since childhood

[DeepSeek @deepseek_ai]

🌐 Supporting open source. Expanding deployment options.

We’ll work closely with the open-source community on V4.1-Flash inference support and explore more deployment options.
Planning a large-scale deployment with 2,000 GPUs + a storage cluster? Let’s talk.

🔹 Model: https://t.co/rv2G2TN66a
🔹 Paper: https://t.co/FIECPM1SSV

6/6

[Didier FRANC @DidierFranc]

@alexsoyes Tu dois avoir un souci, ça devrait pas être aussi lent même sur un M4 Pro. Pour ma part, c'est d'ailleurs beaucoup plus rapide que sur la plupart des providers, où ça devient de plus en plus mou. J'ai un M5 Max.

[Pranit @fazxes]

@DidierFranc coming up - it was more of an internal tool thats why Gateway is basically primary choice but I do agree - Local models are the future.

[Didier FRANC @DidierFranc]

👇

[Didier FRANC @DidierFranc]

Cloud solutions are obvious choices for most people, but many enterprises still rely on isolated, secure environments for their workflows. fx appears to be a promising alternative to increasingly bloated agentic solutions. I look forward to benchmarking it in industrial settings 🙏

[Didier FRANC @DidierFranc]

Cloud solutions are obvious choices for most people, but many enterprises still rely on isolated, secure environments for their workflows. fx appears to be a promising alternative to increasingly bloated agentic solutions. I look forward to benchmarking it in industrial settings 🙏

[Didier FRANC @DidierFranc]

🤯

[antirez @antirez]

Enjoy the DwarfStar glm-5.3-flash branch with GLM 5.3 Flash Q2 and Q4 support: single MacBook 128GB or DGX Spark inference, two MacBook RDMA 128GB each Q4 inference in tensor parallel fashion at 37 t/s single generation. ~500 t/s prefill for now, can be higher. ROCm soon.

[Didier FRANC @DidierFranc]

@mzeller already works on m5 max

[Naval @naval]

It's the Age of Builders.

(sorry financiers and talkers)

[Didier FRANC @DidierFranc]

@ReineCalypso Claude Code → 9router → OpenRouter → Deepseek-v4-flash

[Didier FRANC @DidierFranc]

Following our announcement a few months ago that Agnostic received a grant from the @StellarOrg we’re excited to introduce you to the open Stellar Analytics Platform.

We are launching the first advanced analytics platform designed to make on-chain data truly accessible with real-time dashboards, open SQL access, and an AI-native query engine.

Turning raw blockchain data into actionable insights shouldn’t be hard. Now it isn’t.

Full announcement: https://t.co/WqKBvSXGRz

[Didier FRANC @DidierFranc]

Announcement: Stellar (@StellarOrg) x Agnostic(@agnosticeng)

We’re beyond excited to announce that Agnostic has received a grant from the Stellar Community Fund! 🚀

This recognition is more than just financial support; it’s a strong validation of our vision:
- To make blockchain data instantly accessible, truly real-time, and powered by AI.
- To simplify querying across chains by providing a unified, low-latency analytics layer.
- To enable builders, analysts, and businesses to unlock insights without needing to understand the quirks of every chain.

Until now, Agnostic has been focused on EVM ecosystems. With Stellar, we’re making a bold move:
- Bringing our AI-native query engine to one of the most innovative and mission-driven ecosystems in blockchain.
- Offering real-time dashboards, advanced analytics, and open SQL access for Stellar activity.
- Helping the Stellar community turn raw on-chain data into actionable insights for DeFi, payments, and beyond.

The Stellar Development Foundation and the SCF community recognized our potential, and we’re deeply thankful for their support. 🙏

This is just the start. With Stellar, Agnostic is getting closer to its ultimate goal: to become the AI-powered analytics platform for the multi-chain world.

Stay tuned, the next chapter of blockchain analytics is about to unfold. ✍️

Source: https://t.co/USz3tM27WE

#Stellar #Blockchain #AI #Analytics #DeFi #Agnostic #SCF

[Didier FRANC @DidierFranc]

RT @agnosticeng: 91% performance boost to our indexing!

Thanks to @paradigm's alloy_rs! (https://t.co/wTNsqQpEP7)

Article 👉 https://t.co/…

[Didier FRANC @DidierFranc]

🧠 Want the full deep dive?

Check out the article by @briche_arnaud that breaks it all down:

👉 https://t.co/yfu8Qp1M23

From Go to Rust, UDFs to warp speed — it’s all there. 💥

[Didier FRANC @DidierFranc]

🚨 Speed boost alert!

At @agnosticeng we just doubled our indexing speed on #Ethereum and other #EVM chains by switching from @golang to @rustlang.

The secret sauce? alloy_rs, the little brother of reth, plus a few low-level tweaks.

Blazing-fast, memory-safe, and built for scale. 🔥

[Didier FRANC @DidierFranc]

The results speak for themselves:
⚡️ 2× indexing speed
🛠️ Minimal effort
🚀 Massive gains

Huge thanks to everyone pushing the EVM & Rust ecosystem forward — especially the minds behind reth, alloy-rs, and the broader @paradigm crew. 🫡💙