Patrick Laverty

machinist, tinker, free thinker.

[Beff (e/acc) @beffjezos]

Diffusion of models, private ownership of hardware, and decentralized control of capabilities is the backstop against authoritarian tyranny of centralized institutions merging with the govt and banning AI for everyone else

If they try to ban private AI it's war.

[Patrick Laverty @PatrickLaverty0]

i cant make this shit up

[Paisley Circus @PaisleyCircus]

@peterAjensen Satire? lol

[Chris Dukes @clduab11]

@peterAjensen Not only are you all dead wrong, but I'll be doing what I can to partner with Indiana law firms to enjoin the City of Bloomington from enforcing this law.

You can't even DEFINE "superintelligence", let alone know how it works. See y'all in court.

https://t.co/ZgoQJvLq7x

[Chris Dukes @clduab11]

@peterAjensen If anyone knows any Indiana lawyers; I've drafted a draft Complaint and accompanying Memorandum of Law showcasing why @POTUS should send the Federal authorities to Bloomington, IN. Find the material here for free.

https://t.co/gdt2VpJXv7

[Patrick Laverty @PatrickLaverty0]

@peterAjensen @Chaos2Cured yeah real intelligent reply, a single emoji to that mans nuanced argument. The real clown show starts when you retards think you'll be able to march door to door confiscating peoples hardware. Please send me a notification when you do and live stream it.

[Peng @pengsonal]

4 FREE UNCENSORED AI MODELS THAT RUN ON 8GB RAM. NO GPU NEEDED

run locally with just your CPU + RAM

1. Dolphin 3B → Lightweight chat, ~8GB RAM
https://t.co/Ur3j5HM5rQ

2. Qwen3 4B Abliterated → Fewer refusals, ~8GB RAM
https://t.co/Q3BUDpvplI

3. Qwen3 4B Heretic → Non-thinking chat, ~8GB RAM
https://t.co/6P5xdc46X5

4. Dolphin 8B → Chat + coding, ~16GB RAM
https://t.co/FjWx4EIBa1

setup:

• Install llama.cpp: https://t.co/NQlj1ShsjC
• Download a Q4_K_M GGUF file
• Run: "llama-cli -m model.gguf -ngl 0 -c 2048 -cnv"

replace "model.gguf" with your model's file path.

works offline. CPU generation can be little slow

[Patrick Laverty @PatrickLaverty0]

@maxwinga @ONagel33303 @Gregory_C_Allen The coping is from you retards. You think China is going to stop advancement to appease a bunch of retarded American libshits? You people really are delisional.

[Patrick Laverty @PatrickLaverty0]

lol love to see it. cope antis

[Ori Nagel @ONagel33303]

A bill to Ban Superintelligence is unlikely to pass the 60-vote filibuster in the Senate.

—@Gregory_C_Allen (Former Senior AI Advisor, CSIS)

[CrazyShyyt @CrazyShyyt]

China built a 27B model that matches Claude Fable 5 on coding.

and it's 1000% open-source and free.

[Davey @daveeyyy_]

Jensen Huang: Nvidia already depends on AI to design its chips and write its software.

“We can’t design a chip anymore without AI.”

“At night, our AIs are exploring design spaces, uh, vast and wide, that we would never do ourselves. Because it costs too much money to explore it.”

“We can’t write software without, without AI anymore.”

“I want everybody, every organization in our company to use AI very aggressively.”

“I want, I wanna turn Nvidia into a one giant AI.”

[Li Zexin 李泽欣 @XH_Lee23]

The US that traded slaves and caged living humans as animals in "ZOOS" for amusement are lecturing us barbaric and evil.

US owned Human Zoo.

US voted against UN resolution declaring transatlantic slave trade the gravest crime against humanity in 2026.

[Dolphin Emulator @Dolphin_Emu]

We've released Dolphin 2609b.

This hotfix fixes several issues, such as a bug where Dolphin crashes for users in Alberta and British Columbia, color corruption on Adreno GPUs, missing translations on macOS, and more.

Check here for the full change log: https://t.co/6OCJRK31ue

[Qwen Developers @QwenDevs]

Qwhen? Qsoon!

For our friends outside China, the Chinese text in the poster comes from Zhuangzi, an ancient Chinese classic:

"Far in the northern sea lives a fish named Kun (pronounced like Qwen). So immense is its size that no one knows how many thousands of miles it spans."

[Syd Steyerhart @SydSteyerhart]

You're not afraid that AI is going to end the human race. You're afraid that AI is going to end your bullshit little email job, specifically.

Because it is.

[Dreamcast Nostalgia @Dreamcastalive]

#STARWARS TRILOGY on Dreamcast 🚨**

Yes! Fans can now play this legendary **Model 3 arcade classic**, fully ported to the #Dreamcast ! 🤝🔥

Created by:Toastedgreedo

So, what do you think... does it hold up?😳🔥

Video complete:
https://t.co/GX384YUYCj

[Patrick Laverty @PatrickLaverty0]

@0xCodila thanks, ill just point claude at the repo and tell it to be careful of these tricks

[Patrick Laverty @PatrickLaverty0]

@BAHX3305 Most westerners are too stupid to understand sarcasm. Sarcasm is the highest point of humor, humor is the highest point of human intelligence. You cast your pearls before swine, love.

[Patrick Laverty @PatrickLaverty0]

@PauseAI_UK you cant pause china. cope.

[Bindu Reddy @bindureddy]

🚨 Tons of misinformation are being spread about AI right now

Here are the actual facts

- GPU demand is still sky-high
- token usage is exploding as costs go down
- No, enterprises have NOT hit peak AI
- 99% of workflows still need to be automated
- AI is more correct than almost all humans
- AI helps most humans become 10-100x more creative
- Globally, AI has elevated humans and made the playing field more even

Overall, this leads to a far more efficient, revenue-maximizing system. It will eventually make work optional and usher in the age of abundance

[Nyxelarm @Nyxelarm]

A 2.78 TRILLION PARAMETER MODEL JUST RUNS ON 8GB RAM. NO GPU. NO FRAMEWORKS. PURE C.

Free. Open source. The full Kimi K3. Not a cut-down version.

→ Full 2.78T parameter model
→ Runs on a single CPU in 8.24 GB RAM
→ Pure C99 engine (176 KB). No PyTorch, no CUDA, no BLAS
→ Same output on any machine. More RAM only makes it faster
→ Chat mode. Linux, macOS, Windows

Needs ~1.7 TB of storage and a decent SSD.
As we can see, you need a lot of memory for very big capabilities.

On 8 GB it's slow, but a trillion-parameter model on a normal PC is still wild.

How to get it: Clone the repo, run make, check your machine with ./scripts/k3-doctor.sh, download the checkpoint and pick a preset that fits your RAM.

All parameters and installation details are in the repository link in the comments. 👇