Acquire 🇨🇦

It's official... Picked up my suit 👔 @DoginalDogs DRC-20 #doginals are #ordinals on $DOGE ⛓️ made possible by @VeryDogeLabs Are you going to fade them, ANON? Watch... soon available on @OkX and @MagicEden To the moon 🚀 !woof

@goodvibesclub Quietly owns the room. Goodfella sits at the bar pensively, buzz’d, waiting patiently for a sign that everything will be ok 🤙

Vibetown's hottest new rooftop officially has a name. Welcome to Buzz'd 🍸 Doors are open. Now we need to know who's on the list. Which Citizen from your collection has that rooftop energy? The one who walks in and owns the room. QRT this post with your Citizen and tell us why they belong at Buzz'd 🤙

@DavidOndrej1 Yes, but can you fine-tune MLX on Apple silicon?

this makes fine-tuning 10x easier

@gregisenberg Yes, it is… this is the dream

Is Hermes Agent the new OpenClaw?

@DavidOndrej1 Exciting s*** Thanks for sharing David.

@elonmusk What?! Agents without custom tools calls… 😏

Design your own agents with Grok

@RaminNasibov If you’re looking for a business name Then, logo.

Logo for Polar Bar If you’re looking for a logo nasibov.me/contact

@RaminNasibov Reminds me of Wolfenstein

simple and creative

@Lolagoodvibez Lmao so good 😂

Apparently if you’re against wars you’re a Jew hater. What a time to be alive 😂

@Lolagoodvibez It’s ok… thank you for letting us know. Any interest in Khartoum? 😂

I NEVER had the desire to visit Dubai, ever. Not even the slightest thought.

@tMAIS0N @MasterSANdoteth Good to see you building!

Soon I'm going to show you something I built. Not a skill. Not a workflow. A real product. Built with OpenClaw + Claude Code. I'm excited about this one.

@chooserich Most confusing part is how to design an agent from scratch… what’s the tech stack? Asking for a friend

I’ve spent the last year building with AI. Real workflows. Not hype. Most people are overwhelmed by tools and noise. What’s the most confusing or challenging part of AI for you right now? I’ll break it down clearly.

@frankdegods @tulipking Action. Cognition. Study. Work. Got it… so we routing from memory here?

@tulipking diataxis.fr

there's been a bunch of interesting-ish articles about different systems for agent memory but i'm not really convinced the solution is obsidian-cli or a more opinonated structure of md files those probably net out positive performance gains but my intuition is they're hacky. the real solution is just gonna be 10M+ context windows & some sticker/permanent memory being baked into the models themselves [butchered analogy] we're at the cursor phase of memory implementation (skewmorphic) and we need to get to the claude code phase (native and elegant) of implementation

@EndWokeness Sure… so sane it’s insane 🧠 0.33% (1955) 0.01% (2025) Likely should be closer to 10.8%

The most insane chart you'll see today

@Be_Believing Hilary Clinton top right?!

Antony acted perfectly, one simple mistake and his career would’ve been over..

@ChrisCoffee Interesting, that’s an integration… Tried to whip-up a coding project the other day to generate pptx using this anthropic agent skill 👇 skills.sh/anthropics/ski… Tried extracting styles and assets from old presentations, to give them a refresh ✨

Used Claude Opus 4.6 to build a PowerPoint today. Actually kind of wild how fast it worked. Made a quick how to video if you want to try it out!

@albertadevs So true… got to wait for the dust to settle before making decisions on what your integrating into your tech stack.

You’re not falling behind if you’re overwhelmed by all the new AI coding tools, we’re literally all overwhelmed 😭

@KartikeyStack Thank you! Good timing for a quick drill-down on Redis. Got to break down and separate concerns to address the best data handling strategy.

@jensenloke @AlexFinn Totally local? What open source model would you load-up for orchestration with ollama?

@AlexFinn @AlexFinn is so dam cool. Loled at “stfu” for losers who didn’t get a Mac mini. But the idea is to go local. For what it’s worth I jumped on the bandwagon too!

You should buy a Mac Mini Unless you are EXTREMELY technical I highly recommend AGAINST running ClawdBot on a VPS Despite what all the angry nerds will tell you, Mac Mini is the way When you run Clawd on a Mac Mini, you are secure by default. You have to do stupid things to get hacked When you run Clawd on a VPS, you are exposed by default, and have to do incredibly smart things to become secure On a VPS all your keys, credentials, and access to your life are by default exposed unless you introduce strong security methods that are too technical for the average person to understand On a Mac Mini, your Clawd is by default contained on the device. Unless you expose the Clawd to the world via email, group chats, or other ways you will for the most part (with exceptions, there are no absolutes here) be safer. The Mac Mini (or any local device, old laptop, windows or linux computer) is 1000x better for the average person. You're getting a full time, more secure employee for $600. Incredible deal

@AlexFinn What about those being exposed on those inferences to Anthropic? Only way I’d run clawdmolt is using a local model with ollama for total privacy. If it ain’t Opus 4.5, what open source model would you suggest for a solid experience?

@trikcode Called it.

SaaS is now dead. Welcome to AaaS (AI as a Service)