TheCharlieMachine

1/ While working on BoardGameMatch, I realized I was wasting a stupid amount of time on tiny UI edits.

2/ Every change meant rewriting the same kind of prompt: what to change, where it is, and what code it probably touches.

3/ So the idea is a Chrome extension that lets you draw zones directly on a webpage and auto-generates a Cursor prompt with the relevant code references.

4/ That would make front-end iteration way faster for small visual changes.

5/ Would you use this if you build web apps with Cursor? #Cursor #WebDev #Frontend

@amitisinvesting Noy yet, it needs to fall a bit more to my taste 😊

is anyone listening to Bill Ackman and buying the dip

1. I built a board game matching app that recommends new games based on the ones you already own.

2. Instead of browsing random top lists, you start from your actual taste and your current library.

3. The goal is simple: help board gamers discover better-fit games to add next, faster.

4. Would you use a tool like this for your collection? board game recommendation app, board game matching app, tabletop game discovery #BoardGames #BoardGameApp

How to #motivate your #openclaw 😅 Sometime it just lacks the courage to try things for some reason. But it's opened to #selfimprovement 👍

1. In the West, the loudest AI messaging often comes wrapped in warnings: job loss, extinction, deepfakes, collapse, misinformation, loss of control.

2. In China, AI is more often framed as progress, competitiveness, convenience, education, national strength, and economic upside.

3. People do not just react to technology. They react to the emotional framing around technology. If your elites constantly describe the future as dangerous, people become defensive and pessimistic.

4. Big AI companies may think they are being responsible by emphasizing risks, but constant fear-based projection has second-order consequences: public resistance, cultural pessimism, and weaker adoption.

5. The real question is not whether AI is good or bad. It is who gets to define the emotional story around it — builders, institutions, media, or the public itself.

I tested Qwen 3.5 on an RTX 3090 and honestly... this GPU is still a monster for local AI. Ollama makes the setup stupidly simple, and once you see the tokens/sec vs the VRAM usage, you realize why 24GB still matters in 2026.

2. If you're trying to run serious local models without jumping to server-grade hardware, this is still one of the best sweet spots. RTX 3090 AI, Qwen 3.5 benchmark, Ollama setup, local LLM. #LocalAI #Ollama #Qwen #Hardware

@chamath @KrakenRobotics

“China is conducting a vast undersea mapping and monitoring operation across the Pacific, Indian and Arctic oceans, building detailed knowledge of marine conditions that naval experts say would be crucial for waging submarine warfare against the United States and its allies.” We need to map our oceans. ASAP.

@abalus8 @grok peut-on avoir des détails ?

Les dirigeants de la biotech Carbios visés par une série de plaintes d'actionnaires minoritaires lalettre.fr/fr/entreprises…