Local AI. Building @picogpt Download: https://xitter.cf/t.co/o0PKhVM691 Discord: https://xitter.cf/t.co/KNirjAzl32
San Francisco, CA
Joined April 2008
- Tweets28.8K
- Following6.4K
- Followers6.9K
- Likes104K
Pinned Tweet
Apple Silicon + Gemma 4 fans: this is for you.
Pico AI Server now supports continuous batching with MLX-Swift.
43 tok/s on 1 stream.
26 tok/s per stream on 2 concurrent streams.
That’s 52 tok/s total. a 21% throughput gain on a six-year-old MacBook Pro M1 Max!
Everybody: Please stop posting useless but cool iPhone duo demo. It makes me want to buy one 🤬
STOP IT
⚽️ I’m having WAY too much fun with this game on the iPhone Duo 😂
This is how you take advantage of a foldable!
🥅 The goalie is on the vertical screen
⚽️ You kick the ball using the bottom screen
📱 The entire form factor becomes part of the game!
It genuinely feels like we’re back in the Wild West of apps, when developers made fun, creative experiences just for the sake of having FUN.
We need more of this!! 🔥
I’m still not sure if the vertical controls was a good idea but the open and close transition is a timeless gem (that will be copied by everyone I’m sure)
Ronald Mannak retweeted
No one has ever won both the FIFA Peace Prize and the Citizens of Iowa Man Of The Century award in the same year
Todd gave a great talk at the AiOS meetup this week. If you’re interested in local AI, watch the video
Want to know how on-device AI works on Apple products, from the iPhone all the way up to a cluster of Mac Studios?
I gave a talk at the AiOS meetup in SF on Tuesday. Here’s a video of the same presentation. A lot of people asked for the deck, so it’s up too.
The page credits every source. Special thanks to @alexocheema, whose price-per-GB-of-memory idea I happily stole, to @jundotkim, whose oMLX benchmarks are behind most of the speeds, and to @ashxhart, @MiaAI_lab, @ivanfioravanti and @viticci for their benchmarks, recipes and reviews. And to everyone else listed.
Who are the Bay Area engineers building local AI inference engines or making significant contributions to MLX, ANE, or NVIDIA Spark/CUDA?
I’d love to connect with more of the people behind the code. Who should I follow?
I’ve said it before: while Terminator 3 might be seen by some as a b-movie, it’s extremely underrated (and so much better than whatever came after). Which other movie predicted yolo mode, rogue agents escaping their sandbox and php vulnerabilities?
I had forgotten that in Terminator 3, skynet starts exploiting vulnerable machines everywhere to cause a panic. Skynet agent does not yet have yolo permissions, so they can't activate weapons.
The CIA says "activate skynet to fight it", and JC's father in-law is like "NO I STILL PREFER TO KEEP HUMANS IN THE LOOP" LAMO
OMG. Can someone check on Facebook?
Effective Immediately, all organizations using PHP for their systems are advised to be prepared for an update addressing two critical vulnerabilities. PHP is widely used for web applications. These vulnerabilities often provide a direct path to the server's file system or database. If an immediate update is not feasible, it is recommend that organizations take the system offline until it can be fully mitigated. NSA will follow up shortly with more specific guidance.
I met Ben yesterday. Super bright guy and I would never have guessed it was his first day in SF ever, because he fit in so naturally. Sometimes you just know they’re going to make it in Silicon Valley.
If anyone is working on the intersection of AI and UI, I highly encourage you to reach out and grab a coffee with Ben.
4 days ago I packed a bag and left Ethiopia to come to the U.S.A
Yesterday I landed in San Francisco with a one way ticket. Still determined to change my life in the 3 months that I have here.
I am spending my days going to events and meeting as many devs as I can. I'm spending my nights building my face off.
I couldn't be more excited. Time to chase dreams
