Julian Pscheid

Hedy is now on OpenClaw. 🦞 Install our skill from ClawHub and your AI agents get direct access to meeting transcripts, highlights, todos, and more. Open source. MIT-0 licensed. Three steps to set up. clawhub.ai/JulianPscheid/…

@AlexFinn Thank you sir.

IF YOU'RE ON OPENCLAW DO THIS NOW: I just sped up my OpenClaw by 95% with a single prompt Over the past week my claw has been unbelievably slow. Turns out the output of EVERY cron job gets loaded into context Months of cron outputs sent with every message Do this prompt now: "Check how many session files are in ~/.openclaw/agents/main/sessions/ and how big sessions.json is. If there are thousands of old cron session files bloating it, delete all the old .jsonl files except the main session, then rebuild sessions.json to only reference sessions that still exist on disk." This will delete all the session data around your cron outputs. If you do a ton of cron jobs, this is a tremendous amount of bloat that does not need to be loaded into context and is MAJORLY slowing down your Openclaw If you for some reason want to keep some of this cron session data in memory, then don't have your openclaw delete ALL of them. But for me, I have all the outputs automatically save to a Convex database anyway, so there was no reason to keep it all in context. Instantly sped up my OpenClaw from unusable to lightning quick

@mattsilv Good thing my lobster runs on GPT

@JulianPscheid At least yesterday we could sonnet max

they can't keep getting away with this

@AlexFinn I already have the Max subscriptions on both Claude and Codex, so might as well just use those.

I don't care what computer you have, you should be running local models It will save you a money on OpenClaw and keep your data private Even if you're on the cheapest Mac Mini you can be doing this Here's a complete guide: 1. Download LMStudio 2. Go to your OpenClaw and say what kind of hardware you have (computer and memory and storage) 3. Ask what's the biggest local model you can run on there 4. Ask 'based on what you know about me, what workflows could this open model replace?' 5. Have OpenClaw walk you through downloading the model in LM Studio and setting up the API 6. Ask OpenClaw to start using the new API Boom you're good to go. You just saved money by using local models, have an AI model that is COMPLETELY private and secure on your own device, did something advanced that 99% of people have never done, and have entered the future. There are some amazing local models out there too right now. Nemotron 3 and Qwen 3.5 are fantastic and can be ran on smaller devices Own your intelligence.

@AlexFinn Also great if your house is cold and you just want to have a nonstop heater on your desk.

@mattsilv @heygurisingh 🚨 BREAKING: I am an AI engagement farmer. Look at my shit.

@heygurisingh this is cool, but please don't call every cool thing breaking news. this project has been around for many months. hope to see you again soon

🚨BREAKING: This open source tool generates entire Minecraft worlds from real-world map data. It's called Arnis. You pick any location on Earth. Arnis builds it block by block in Minecraft. No mods. No world editors. No manual building. Just real geography. Here's how it works: → You select any area on the map using a rectangle tool → It pulls geospatial data from OpenStreetMap → Processes real-world elevation and terrain data → Maps every building, road, river, and landscape to Minecraft blocks → Generates a fully playable Java or Bedrock Edition world Here's the wildest part: It doesn't just flatten everything into a 2D grid. It reconstructs actual topography. Hills, valleys, coastlines. Real elevation data turned into Minecraft terrain. You can generate your hometown. Tokyo. Manhattan. The Swiss Alps. Anything OpenStreetMap has data for. Tom's Hardware, Hackaday, and XDA Developers all covered it. It even got cited in an academic paper on using Minecraft for flood preparedness education. Written in Rust. Ships with a full GUI. Works on Windows, macOS, and Linux. 32 contributors. 746 forks. 8,100+ GitHub stars. 100% Open Source.

@jeiting OpenClaw also makes for a great product ops person

@JulianPscheid OpenClaw?

Six weeks ago, I fell down the OpenClaw rabbit hole. One of my experiments was to give it an API key for one of my RevenueCat apps, and I was blown away. It made things that would take minutes in the dashboard take seconds. With no handholding. Immediately I started productionizing it, building my first production feature for RevenueCat in 5 years (with some help from some friends). Using codex + claude code, we quickly built up a prototype which is in private beta right now: Rico, our first in-house agent. revenuecat.com/feature/ai-age… It's clear to me now that CRUD software won't be enough anymore. Many interactions with software that were once tedious become magical if you just pour some tokens on it. We will need to reinvent much of what it means to be a SaaS. It will require new thinking, and RevenueCat needs to get there fast. We now have several agentic features in the works, and I think it's just the beginning. RevenueCat of today is the brain stem, AI and agents are the neocortex we will layer on top to create a truly powerful, thinking machine that will help any developer make more money, autonomously. As part of that we need help. We're catching up quickly here, learning how to build and deploy agentic systems. But we need to build a team, and I'm looking for the first product engineer to own how intelligence is built into RevenueCat. You'll work directly with me, @elwatto and others to bring the vision to life. Apply, or send me a DM with the craziest AI system you've built jobs.ashbyhq.com/revenuecat/ec6…

@jeiting Sure is

@stevibe This looks too familiar.

I'm obsessed with pushing local small models to their limits. Qwen3.5:0.8b doing real-time video captioning on a Mac Studio M2 Ultra, streaming descriptions as the video plays. Under 1s per frame — 269 frames captured & described from a 3m49s video. Pause anywhere and read the captions, it describes every frame surprisingly well. This model is barely 1GB. Local AI is moving absurdly fast.

@mattsilv @Polymarket Or shorting himself

@Polymarket Marco Rubio betting on himself on @Polymarket

BREAKING: Marco Rubio closes in on Gavin Newsom for the 2028 Presidential Election.

@jeiting apps.apple.com/us/app/hedy-ai…

Are you an RC customer and want your app icon featured on our signup page? We're building an app icon wall ala WWDC 2009(?) and looking for volunteers. Reply with your app link and you hereby officially consent to me putting in on our website (this is legally official)

@mattsilv yup, trying the codex app (have only been using the CLI so far) and it's pretty damn good

@JulianPscheid or to try antigravity/codex osx app

anyone else unable to authenticate with claude code CLI today? I I tried restarting it on two different Claude accounts I have: one for work and one for personal, and I cannot use the OAuth login with Claude code CLI this morn

Oh no... a Codex CLI outage has forced me to try the Codex desktop app... and I must say... it's PRETTY DAMN GOOD 😳

@ChrisKruegerDev @RevenueCat 😮

Pretty nice to see the ad revenue alongside the in-app subscription revenue in @RevenueCat.

@PaulHShapiro @theEVERYcompany Amazing!!

Real egg white protein powder, made from microbes, not chickens. What a cool future to be living in. 🚀 😎 Congrats to @theEVERYcompany, a pioneer in the precision fermentation space, on this launch. (Yes, also I haven’t shaven for a little while.)

@HedyAI_ 👏

Hedy 2.15 is live. European users can now store their conversation data on EU servers. Existing users can route AI processing through EU infrastructure from Account Settings. Also new: Negotiation session type with real-time coaching 120-min automatic suggestions (up from 60)