AI Tool of the Day: Meshy by @MeshyAI. Turn a prompt, an image or a few phone photos into a textured 3D model, then rig it, animate it or get it ready for a 3D printer. Free plan with 100 credits a month; paid plans from $20 a month. Models are made on Meshy's servers, not on your Mac.
2
4
537
Price, licence and where your data goes, each with a source, in our catalogue: apps.mlx.app/app/meshy?utm_s…
2
388
for scale: 1M output tokens from a 35B MoE on a base mac mini (~40 tok/s) is about 7 hours and roughly 7 cents of electricity. it won't port Battlefield as well as Claude does yet, but that gap is closing faster than the price is
Decompiling Battlefield 1942 and then porting it to C and then WASM cost me 1 million output tokens
At API prices that's about $100
But I pay $100/mo Claude
9
463
replies under every mac rumour: "waiting for the M7"
replies from the local AI crowd: "cool but how much memory"
2
253
He's right about the 5%.
We'd just rather put the other 95% to work than move it all into a data centre.
WAIT… I can’t believe Sam Altman really said this out loud about local AI: “A lot of people want local inference. And they assume that the problem is solved now because we are going to have open source models. And so we're not going to need all the stuff in the data center.
But from a global allocation perspective, actually, it's quite terrible to have everybody have a bunch of the world's limited supply of DRAM sitting in their own computer. And, being used 5% of the time.”
this is such a bizarre thing to say in the open. I don’t get why people buying hardware for their own use is framed as a competition with data centers spending billions to buy hardware at scale.
if anyone should be complaining about how the world’s memory is allocated, it must be the average consumers facing the skyrocketing hardware prices.
it just doesn’t make any sense.
1
4
1,093
MLX Node lets a Mac lend its idle hours to an AI network, within limits its owner sets. Why idle Macs make sense as compute:
mlx.app/learn/apple-minions-…
3
585
He kept the box.
You kept the brains.
MLX Node runs network test inference on an awake, plugged-in Mac. Your hours. Your limits.
Download free → mlx.app/download?utm_source=…
🤖 Made with AI
3
358
Brush. Charge. Press Start.
Stop whenever you choose.
Short 10/10 — Ritual.
MLX Node runs network test inference on an awake, plugged-in Mac. Your hours. Your limits.
Download free → mlx.app/download?utm_source=…
🤖 Made with AI
3
289
His biggest fan has four legs.
Short 09/10 — The Final.
A fictional match against doing nothing.
MLX Node runs network test inference on an awake, plugged-in Mac. You set the limits and choose when to start.
Download free → mlx.app/download?utm_source=…
🤖 Made with AI
2
248
Infrastructure. Meet houseplant.
Short 08/10 — Small Talk.
A fictional conversation between a data center and a Mac.
Put spare Mac time to work with MLX Node. Network test inference; you choose when it runs.
Download free → mlx.app/download?utm_source=…
🤖 Made with AI
3
206
For once, doing nothing made the news.
Short 07/10 — Breaking News. A fictional bulletin for very real idle Macs.
Local chat, coding agents and APIs: MLX Harness → harness.mlx.app/?utm_source=…
Network test inference: MLX Node → mlx.app/download?utm_source=…
🤖 Made with AI
4
176
Before downloading a local model, check runtime support for that exact architecture/format and memory for weights, context and other apps. Weight size alone isn't the memory bill.
5
142
Make your first local API call boring.
In MLX Harness:
1. Choose a model in Chat & Images; send a short test message.
2. Open Servers & API → API. Copy the app’s curl snippet.
3. Run it unchanged before wiring up your agent.
OpenAI-compatible base URL: http://localhost:8080/v1
Use the model name in the snippet.
One working request. Then the clever stuff.
Download free → harness.mlx.app/?utm_source=…
1
5
152
Clock out. Leave the Mac clocked in.
Short 06/10 — Lights.
MLX Node runs network test inference on an awake, plugged-in Mac. You choose its hours and limits.
Download free → mlx.app/download?utm_source=…
🤖 Made with AI
1
5
131
Leave lazy to the expert.
Your Mac has other talents.
MLX Node. Download free → mlx.app/download?utm_source=…
🤖 Made with AI
4
108
Previous experience: spreadsheets.
Reason for leaving: bored.
Short 05/10 — The Interview.
MLX Node runs network test inference on your Mac. Set your limits. Start or stop when you choose.
Download free → mlx.app/download?utm_source=…
🤖 Made with AI
3
96
Guilty of excessive potential.
Short 04/10 — The People v. The Idle Mac.
Run your own local AI: MLX Harness → harness.mlx.app/?utm_source=…
Contribute spare Mac time: MLX Node → mlx.app/download?utm_source=…
🤖 Made with AI
4
84
Night shift.
No paranormal experience required.
Short 03/10 — 3:07 A.M.
MLX Node can run while you sleep. Your Mac stays awake and plugged in; you set the limits and choose when to start.
Download free → mlx.app/download?utm_source=…
🤖 Made with AI
4
76