Joe Bighill

@NVIDIA_AI_PC 👀

Be honest — how many local models do you have downloaded right now? 👀

Do… do people following AI really, honestly, genuinely not know who “Claude” is named after!?

Saw this over on Nostr. Haven't seen it here yet.

Bitcoin is about to shock the world

???

Claude Code is now working full-time keeping my @openclaw alive and functioning, with several hacks and tricks. And poor GPT-5.5 is oblivious as to why it is not working.

@coinbureau M4 Ultra?

🚨OPENCLAW PUSHES MAC MINI INTO AI SPOTLIGHT The $599 Mac mini has become one of the hottest machines for local AI development after OpenClaw helped make Apple’s unified memory architecture ideal for running large models. Apple’s M4 Ultra supports up to 192GB unified memory, far above the ~32GB VRAM on consumer Nvidia GPUs. Apple CEO Tim Cook says Mac mini and Mac Studio could remain supply-constrained for months as AI demand exceeded forecasts.

you can outsource your thinking but you cannot outsource your understanding

@ivanfioravanti Thinking on?

For the time being Qwen3.6 27B is still my preferred local model.

This is bad, and is only going to get worse as the staring point of the calculation gets more red.

Correction: Thanks to Wolf Of All Streets. "June 2022. Bitcoin bottomed at 17,600 and made a move to 25,200. Roughly a 43% gain. Then it went down to 15,500 in November." There was a slightly lower low of 10%, equivalent to Bitcoin going from 60K to 54K. I missed that one.

Say woopsie without saying woopsie.

@Shpigford Anthropic just handed 40+ companies $100M in compute credits to run Mythos. When each session burns up to 50M tokens, the math isn’t hard. Paying subscribers are subsidizing Project Glasswing.

i don't believe this uptime for one second. not a day goes by that i don't hit 500 errors. i'd be shocked if they truly even have ONE nine of uptime. i get it. unprecedented scale, etc etc. but i'm struggling to find a service i trust less to actually work. /end rant

Five Claude outages in the first eight days of April. Mythos isn’t released, but Anthropic just handed 40+ companies $100M in compute credits to run it. When each Mythos session burns up to 50M tokens, the math isn’t hard. Paying subscribers are subsidizing Project Glasswing.

@bridgemindai GLM 5.1 has max concurrency=1 Decent model though, but the 429s from z.ai is getting frustrating really fast.

GLM 5 Turbo and GLM 5.1 are ranked #2 and #4 on DesignArena. Sandwiching Claude Opus 4.6. A model that costs $1.40/$4.40 per million tokens is producing Claude Opus 4.6 level design output. The $200/month model and the $80/month model are neck and neck on code design. GPT 5.4 is nowhere near the top. Gemini 3.1 Pro is mid. Grok 4.20 near the bottom. GLM is not benchmaxed on design. This is real.

THIS IS ACTUALLY THE PHOTO OF ALL TIME HOLY SHIT

Let me explain exactly why Apple still uses drag-to-install in 2026, because the joke here accidentally proves Apple right. A macOS .app is a single self-contained folder disguised as a file. Every dependency, every framework, every resource lives inside it. Drag it to Applications, it works. Drag it to Trash, it's gone. No registry entries. No leftover DLLs. No uninstaller that misses half the files. Windows installers scatter fragments across Program Files, AppData, the registry, system32, and a dozen temp directories. Uninstalling a Windows app is an archaeological dig. Five years later you're still finding config files from software you forgot you owned. Linux is worse. Dependency hell is so common they named it. Entire package managers exist to solve the problem of "I installed something and now nothing else works." Flatpak and Snap were invented specifically to copy what macOS bundles already did natively. The macOS bundle architecture came from NeXTSTEP in 1989. Steve Jobs brought it to OS X in 2001. The core design hasn't changed because the core design was correct. An app is a folder. Installation is a copy. Removal is a delete. Three operations that map perfectly to how humans already think about files. The drag-to-install window with the arrow isn't lazy UX. It's the entire thesis of the system made visible. You are literally just moving a folder. There is no "installation" step because there's nothing to install. The app is already complete. Every other OS eventually tried to get here. Windows got MSIX. Linux got Flatpak. Mobile figured it out from day one because phones shipped after Apple proved the model. The pattern everyone else converged toward is the pattern this tweet is calling outdated. The funniest part: the app being dragged in that screenshot is Claude. An AI that can write code, analyze documents, and reason about complex systems. And the most advanced step in getting it onto your machine is holding down a mouse button and moving your wrist two inches to the right. That's not a design failure. That's a 37-year-old architecture so good that the most sophisticated software on earth still ships inside it.

@garrytan We think this is an overactive abuse-detection system, digging in. Also working on making the terms for -p crystal clear.

@ashen_one Try /think xhigh inline in chat. You get a noticeably better agent; and alas, twice the strain on usage.

been on the $20/m sub for openai on one of my openclaws and that thing does NOT run out like hes been cooking for a whole day and is still going $20/m claude would've died 6 hours in imo (its still so much dumber tho even on 5.4)

🦀 Big Anthropic debacle this week: They’ve blocked Claude subscribers from using their paid accounts with third-party harnesses. Switched to Kimi-Code as orchestrator in OpenClaw — and it’s already showing real promise. It needs to be reminded to read skills in every new chat, but once guided it performs strong. Still needs some training wheels to fully match Sonnet, but the potential is clearly there. Open-source orchestration just got way more interesting. Who else is testing? #OpenClaw #KimiCode