MUHAMMAD MANNIR AHMAD

Computer Systems Engineer

[Tom Dörr @tom_doerr]

Generates high-fidelity 3D facial animation from an audio source with accurate lip-sync and nuanced emotional expression.

https://t.co/JhPGY2upsj

[ollama @ollama]

You can use OpenCode Desktop with Ollama!

Try it with the top open models!

[How To Prompt @HowToPrompt__]

The entire RAG industry is about to get cooked by… mp4 files!

it's called Memvid and it packages your entire agent memory, data, embeddings, search index, metadata, into a mp4 files.. no database needed. no server. no .wal, .lock, or sidecar files. ever.

It replaces Pinecone, Weaviate, ChromaDB, and every vector database on the marke

Here's what it does:

→ +35% better than SOTA on long-term memory benchmarks
→ +76% better multi-hop reasoning
→ 1,372× faster than standard RAG pipelines
→ 0.025ms latency (yes, microseconds)
→ Rewinds, replays, and branches memory like git
→ Works completely offline
→ Ships in Python, Node, Rust, and CLI

memory is append-only, like git for your agent's brain. you can rewind, replay, or branch any past state. time-travel debug what your agent knew, when, and why.

works offline. model-agnostic. SDKs for Python, Node, Rust + a CLI.

15.7k stars. 100% open source.

[Google Gemma @googlegemma]

Gemma 4 is live on @cerebras, the fastest multimodal inference ever!

Running on Gemma 4 31B open-weight model at a blistering 1,500+ tokens/sec. That's a 15x speedup, unlocking real-time visual and agentic loops without the GPU lag.

[Johannes Maria @luo_yuehan]

China's Al strategy: Ten times less spending, winning the customer race. The secret?
Distillation.
#Al #China

[How To Prompt @HowToPrompt__]

someone open-sourced 15TB of physics simulations that would take a national lab and millions in supercomputer time to reproduce.

it's called The Well, a massive collection of 16 datasets covering:

→ turbulent fluid dynamics
→ magneto-hydrodynamic extra-galactic simulations
→ supernova explosions
→ acoustic scattering
→ active matter (biological systems)
→ 11 more physical domains

backed by Flatiron Institute + 11 universities (Cambridge, Princeton, NYU, Tokyo, Berkeley, Los Alamos).

every dataset is built to train PDE surrogate models, the neural nets that replace million-dollar physics solvers with a single forward pass.

running these simulations yourself needs weeks of supercomputer time and grant-funded HPC access most researchers will never touch.

this is 16 physics domains' worth of that, already done, ready to drop into a DataLoader.

100% Open Source.

[Tom Dörr @tom_doerr]

Open-source face recognition system with landmark, mask, and demographic detection.

https://t.co/1hHpuIHPcg

[Elon Musk @elonmusk]

Based on strong positive feedback from customers in our beta test program, @SpaceXAI will make Grok 4.5 available to the public tomorrow.

It is an Opus-class model, but faster, more token-efficient and lower cost.

[CopyRebeldia @CopyRebeldia]

La biología en PDF acaba de morir otra vez.

Un tío hizo una app donde rotas células, aíslas orgánulos y comparas estructuras 3D como un videojuego.

UI: GPT Images 2. Código: Gemini 3.1 Pro.

Los libros de texto ya no mandan.

[Google Gemma @googlegemma]

Gemma 4 now works on-device using React Native!

You can now run Gemma 4 fully offline in your cross-platform apps with local hardware acceleration:
⚡ Vulkan delegate on Android
⚡ MLX delegate on Apple Silicon

See Gemma 4's vision and tool-use capabilities in action, instantly reading a flyer and scheduling a calendar event, 100% on-device.

[Nous Research @NousResearch]

Hermes Agent is now in the Cloud!

Setup couldn't be simpler: pick a model and a server size. Two clicks and 60 seconds later, your agent is live.

Running a team? Spin up agents for everyone at your org with granular access controls and unified billing, all from Nous Portal.

[Nous Research @NousResearch]

Hy3, the new 295B MoE model from @TencentHunyuan, is now free in Nous Portal for the next two weeks!

It is focused on cost-effective agentic use, and particularly strong on coding, tool-calling reliability, reasoning, and 256K long-context tasks.

[Kshitij Mishra | AI & Tech @DAIEvolutionHub]

WEB SCRAPING JUST GOT A SERIOUS UPGRADE.

PixelRAG doesn't read HTML. It reads the page exactly like you do.

100% open-source.

Instead of parsing websites into plain text, it captures screenshots and lets a vision model retrieve answers directly from the pixels.

Why that's a big deal:

• HTML parsers silently lose information.
• Tables, charts, formulas, and layouts often disappear.
• Even changing the parser can swing RAG accuracy by ~10%.

PixelRAG skips that entire bottleneck.

It indexes what users actually see.

The team built a visual index of 30M+ Wikipedia screenshots, and it outperformed the strongest text-based RAG baseline by 18.1% on text-only QA.

Even cooler:

It includes a Claude Code plugin that gives Claude visual browsing.

Instead of scraping the DOM, Claude can screenshot any webpage, PDF, arXiv paper, or even your local app—and answer based on the rendered page.

The pipeline is surprisingly clean:

→ Render pages into image tiles
→ Embed with Qwen3-VL-Embedding (LoRA-tuned on screenshots)
→ Store in a FAISS index
→ Search visually

The best part?

Upgrade to a better vision model later, and you don't need to rebuild the index.

Because the index stores pixels, not parsed text.

Fully open-source under Apache 2.0.

GitHub: https://t.co/B7whbNg60s

[How To Prompt @HowToPrompt__]

China has killed the entire vector database industry.

They open-sourced TencentDB Agent Memory. It gives any AI agent long-term memory that runs 100% locally.

No Pinecone. No cloud APIs. No repeating yourself every session.

- 61% fewer tokens
- PersonaMem accuracy: 48% → 76%
- Zero external API dependencies
- Runs on plain SQLite

Most memory systems compress your history into an opaque vector pile. when recall goes wrong, you're guessing. this one doesn't compress, it builds a semantic pyramid.

L0 Conversation → L1 Atom → L2 Scenario → L3 Persona.

Short-term state gets encoded as a Mermaid graph in your agent's context. verbose tool logs get offloaded to disk. when the agent needs proof, it drills back via node_id to the exact raw log.

no lossy compression. every layer is readable markdown you can just open and inspect.

5.1k stars. 100% Open Source.

[Rossst.03 @Rossst_03]

Alan Oppenheim, MIT professor:

"Every trader alive uses a moving average. Almost none of them know it's a filter, and funds pay quants $600K for the version they were never taught."

a moving average is a filter. it smooths a jumpy price to reveal the slower signal hiding underneath. that is all it is, and it is the bluntest one there is. the same math gives you filters that lag less and cut more noise, and that gap is a real edge.

this is the second half of what Simons was doing. first you accept the market is a signal buried in noise. then you build the filter that reads it, and the moving average on your chart is the crudest possible version. Oppenheim builds the real ones at the board in this lecture.

here is the catch. a filter only helps if there is a real signal under the noise. smooth pure randomness and you will see a beautiful trend that means nothing. the filter is free. knowing there is a real signal worth filtering is the edge.

[Matt Dancho (Business Science) @mdancho84]

RIP document extractors.

Google just released LangExtract: Open-source. Free. Better than $100K enterprise tools.

Here’s what it does: 🧵

[Tansu Yegen @TansuYegen]

Most laptops feel slow because of one fix you keep ignoring.

[Tom Dörr @tom_doerr]

Builds real-time interactive 2D and 3D web applications

https://t.co/hE8BR3Cf61

[Google Gemma @googlegemma]

Gemma 4 31B at over 1,800 tokens per second!

Gemma 4 is now in Public Preview on Cerebras.

[OpenClaw🦞 @openclaw]

OpenClaw is now on iOS + Android 🦞

📱 Native mobile apps, finally
💬 Agents in your pocket
🔔 Channels, tasks, replies on the go

Run agents from wherever your thumbs are.

iOS: https://t.co/7LHHc9htgM
Android: https://t.co/X0Wuh2uA8w