Kim jensen

[Nicolas Camara @nickscamara_]

we built pdf-inspector so agents can process PDFs without waiting on OCR. it classifies any PDF in ~20ms and extracts clean markdown locally

→ 200 PDFs processed in 2.8s
→ top quality in extracting tables + graphs
→ built in rust
→ open source

https://t.co/Wjp9kpTHXJ

[0xMarioNawfal @RoundtableSpace]

Someone open-sourced GeoLibre, a free DuckDB-WASM 3D spatial mapping engine that runs GIS analytics for Earth and space directly in your browser.

[0xMarioNawfal @RoundtableSpace]

Someone used Claude Code with Opus 5 to build SNOWFLOW, a browser WebGPU snow and waterbending demo with persistent terrain deformation in 9 hours.

- Full-Stack Graphics Pipeline: Synthesized custom WGSL shaders, WebGPU pipelines, and Babylon.js terrain logic from scratch.
- Interactive Terrain Deformation: Renders persistent snow trails, berms, and water-spell displacement that heal dynamically over time.
- Visual Self-Correction: Opus 5 iterated on visual polish and shader bugs directly from browser screenshots.
- 9-Hour Build Window: Delivered a production-ready WebGPU tech demo in 9 hours using ~4 million tokens.

[Kim jensen @PIRealityio101]

@isofunds @EXM7777 Which ones? Because this doesn’t look like it is open source and Can be runned locally

[Fish Audio @FishAudio]

Today we’ve raised $52M Seed and we are announcing the public launch of S2.1 Pro.

>It can clone a voice from 5 seconds of audio
>2x faster than Cartesia & 1/6th the cost of Eleven Labs
>most expressive model with word level control over emotion, intonation, pacing etc

We support frontier AI companies including HeyGen, LiveKit, Retell, Sanas, and OpenArt all run our model in production.

If you're a business and we can't cut your voice AI costs by 50%, we'll give you 1 year of Fish Audio for free.
Book a demo: https://t.co/vHkyZf9JoG

To celebrate our first birthday, we'll give you 1 month of S2.1 Pro for free. Like, retweet, and comment “Fish” to get it.

[Yuma @yumahey]

Added choke point simulation.

You can now close choke points and simulate what it would do with the rest of the trade flows globally.

Also trade flows now up to date to march 2026.

[Kim jensen @PIRealityio101]

@kurorosage How do you bookmark, can’t really understand that part

[pulmencr @pulmencr]

A guy just turned standard home Wi-Fi into a radar that sees through walls.

Over 62,000 stars on GitHub for a system that tracks humans without using a single camera lens.

It doesn’t just know you’re there.

Watch what happens when a person lies down on a bed in another room.

The AI instantly maps their exact posture, tracks their breathing, and calculates their heartbeat in real-time.

No optical surveillance, no privacy leaks - it just analyzes how your body disrupts ordinary radio waves.

The whole thing runs on a tiny $5 microchip.

Traditional security requires a direct line of sight and invades your personal space.

Now, your Wi-Fi router is a camera that sees through solid brick.

[Nick @NickDevFE]

img2threejs - turn one object photo into a code-only procedural Three.js model

Open-source toolkit that rebuilds the object in a single reference image as procedural Three.js code (no mesh downloads), quality-gated by a render-vs-reference loop - strong for hard-surface objects

[Sarah Chieng @MilksandMatcha]

We're giving away 20 Codex Pro plans.

The first 20 U.S./D.C. participants age 18+ to complete our FREE 1.5 hour Deep Learning course by July 20 will receive a free 1 month Codex Pro plan. No purchase necessary. OpenAI terms apply. While supplies last.

Comment or send us your certificate.

[ThreeJS Assets @threejsassets]

New Three.js City Asset Pack is out now!

🏙️ 66 assets — towers, shops, street furniture, a full vehicle fleet
🎨 4 palettes: Day, Warm (golden hour), Cool (overcast), NIGHT
📦 Combined Draco catalog GLB — 184 KB

Get the All-Access Pass or get the single pack on our website 🔥

[Superman @thesupermanmx]

China open-sourced a model that reconstructs any scene in 3D from a regular video, in real-time.

one camera. no LiDAR. 10,000+ frames without falling apart.

just walk around with your camera and watch the entire world get rebuilt in 3D at 20 fps.

→ runs at ~20 FPS on a single GPU
→ Stable over 10,000+ frames
→ Beats optimization-based methods on benchmarks
→ Works on drone footage, driving videos, indoor walkthroughs

100% open source.

[Kim jensen @PIRealityio101]

@thsottiaux No 5H limit, but add a Weekly reset… maybe not for 100%; but for 20, so that people don’t get stuck in the middle of a run and have to have days

[Kim jensen @PIRealityio101]

@strimblez A simulation wargame for geopolitical risk assesments

[Kappaemme @Kappaemme1926]

CODEX SKILL THAT FINDS COMPLEXITY HOTSPOTS IN YOUR CODEBASE!

I made a Codex skill that finds performance hotspots, proposes one safe optimization, and proves whether it actually worked.

Scan your project while Codex creates a repeatable benchmark, asks for approval before changing any files, runs the tests, and compares real before/after performance.

→ performance hotspot detection
→ mandatory approval before edits
→ repeatable before/after benchmarks
→ median, p95 + noise analysis
→ PROVEN / INCONCLUSIVE / REGRESSION verdict
→ behavior and test verification
→ native Markdown report
→ one-command install

Install:
npx --yes codex-speedup-proof

100% open source.
Repo in bio.

[Vaibhav (VB) Srivastav @reach_vb]

codex tip: ask codex to do its research first and then use set_goal to set an appropriate goal instead of /goal

It results in massively better prompt and downstream results!

all my prompts would start with requirements gathering and research and once done set_goal

[Thijs @cdngdev]

i gave 5.6 sol access to my camera roll and had it extract pictures of every piece of clothing i own from my photos

then, told it to find new outfits for me and render them on me with gpt-image!

its kinda cool to see your entire wardrobe in a collection like this

[codila @0xCodila]

Anthropic and Andrew Ng built an agent that uses 90% fewer tokens from scratch:

they dropped the entire book of Frankenstein into a prompt - 108,000 tokens asked one question

the input dropped from 108,000 tokens to 11

here's how:

step 1 → put everything that never changes at the top - tools, then system, then docs

step 2 → mark where the static part ends. everything above it gets cached

step 3 → one stray space breaks it - and you pay full price again

step 4 → the cache dies in 5 min - every read resets the clock

step 5 → cached tokens don't count against your rate limits. free headroom

most people never touch this - it pays for itself on day one

watch & bookmark - this 1-hour brilliant course ↓

[witcheer @witcheer]

few days ago I asked how you run Hermes.

250+ replies came back: 70 of you gave real answers and 60 described full setups.

I read and hand-tallied every one. here is what this community runs, in six numbers. 🧵

[Elon Musk @elonmusk]

Literally true