Loading video...

Video Failed to Load

Go Home

Introducing MLX-Swift-TS An SDK for running time series foundation models fully on-device on Apple Silicon. When I joined Datadog, Inc. , I was introduced to Toto, our time series foundation model, and got excited about zero-shot forecasting across different domains. While building a health copilot app, I realized there...

kunal732's profile picture

Kunal Batra

1,784 subscribers

12,575 views • 4 months ago •via X (Twitter)

LIVE

Anya Rossi• Live Now

Private livecam show

0 Comments

No comments available

Comments from the original post will appear here

Related Videos

Today I'm launching a very early release of MLX Model Manager ( A Swift package for quickly and easily adding LLM/VLMs with just a couple of lines of code into your MacOS/iOS applications for local and private inferencing. This is built on the amazing work of the MLX team & what they have done with MLX Model Manager unifies the work they have done with MLXLLM as well as MLXVLM into one package with added abstractions to simplify the usage. I've also added support for Google's new Paligemma2 Vision Model ( You can see a demonstration in the video below! Huge shoutout to Awni Hannun for the creation of MLX, David Koski on leading the charge for MLXVLM and Prince Canuma for the creation of MLX-VLM. Three massive pillars in the MLX community among many others.
0:11

Today I'm launching a very early release of MLX Model Manager ( A Swift package for quickly and easily adding LLM/VLMs with just a couple of lines of code into your MacOS/iOS applications for local and private inferencing. This is built on the amazing work of the MLX team & what they have done with MLX Model Manager unifies the work they have done with MLXLLM as well as MLXVLM into one package with added abstractions to simplify the usage. I've also added support for Google's new Paligemma2 Vision Model ( You can see a demonstration in the video below! Huge shoutout to Awni Hannun for the creation of MLX, David Koski on leading the charge for MLXVLM and Prince Canuma for the creation of MLX-VLM. Three massive pillars in the MLX community among many others.

Kunal Batra

104,815 views • 1 year ago

Qwen 3.5 0.8B, Gated DeltaNet attention is running on Apple Neural Engine ~56 t/s in LUT6 quantization with some room for optimization left. It is CoreML, Swift and IOSurface on M4Pro. It will slow down as we increase context, but not by much. I think Private API opens the way to integrate ANE with GPU/MLX and possibly some MoE.
0:19

Qwen 3.5 0.8B, Gated DeltaNet attention is running on Apple Neural Engine ~56 t/s in LUT6 quantization with some room for optimization left. It is CoreML, Swift and IOSurface on M4Pro. It will slow down as we increase context, but not by much. I think Private API opens the way to integrate ANE with GPU/MLX and possibly some MoE.

Anemll

13,589 views • 4 months ago

made something I wanted for a while: a high performance custom tab/splits library in swift for native Mac apps. with support for drag/drop, CoreAnimation accelerated transitions, and more. spent a lot of time on the polish of this thing to feel snappy, but fun.
0:18

made something I wanted for a while: a high performance custom tab/splits library in swift for native Mac apps. with support for drag/drop, CoreAnimation accelerated transitions, and more. spent a lot of time on the polish of this thing to feel snappy, but fun.

almonk

25,165 views • 6 months ago

“They asked if I would mind if Taylor Swift did this thing, and I said, ‘No one’s ever going to say no to her,’ we had an hour, maybe an hour and a half [to rehearse]. She knew the song. There wasn’t anything to do, I said, ‘Do you want to do it in another key?’ But she said no.” — Randy Newman spoke to Variety on his duet with Taylor Swift for “You’ve Got A Friend In Me” at the ‘Toy Story 5’ premiere. 🥹
0:18

“They asked if I would mind if Taylor Swift did this thing, and I said, ‘No one’s ever going to say no to her,’ we had an hour, maybe an hour and a half [to rehearse]. She knew the song. There wasn’t anything to do, I said, ‘Do you want to do it in another key?’ But she said no.” — Randy Newman spoke to Variety on his duet with Taylor Swift for “You’ve Got A Friend In Me” at the ‘Toy Story 5’ premiere. 🥹

The Taylor Swift Updates

151,883 views • 27 days ago

Late night gym and swim. Fit. Healthy. Confident. Steadfast. My swift responses to Irelands issues and needs has been a direct rocket each time. You have fully agreed with all my points. No one on this ballot has what I have, nor the ability to do what I do, for the people and land of Ireland, of which I hold so dear! McGregor as Irelands President is A LOCK! 🔒
0:11

Late night gym and swim. Fit. Healthy. Confident. Steadfast. My swift responses to Irelands issues and needs has been a direct rocket each time. You have fully agreed with all my points. No one on this ballot has what I have, nor the ability to do what I do, for the people and land of Ireland, of which I hold so dear! McGregor as Irelands President is A LOCK! 🔒

Conor McGregor

386,333 views • 11 months ago

Last year, almost to the day, Kayla Nicole went on Angel Reese podcast and talked about getting dumped by an NBA player and dressing up as the girl she got dumped for for Halloween. That was Teyana Taylor that she was shading. A week ago, Taylor Swift went on Jimmy Fallon and said the last movie she watched was the one starring the one and only, Teyana Taylor lol. She praised all the actors. Taylor swift is a GD mastermind and I love it. I have seen NO ONE catch this…. If it starts going viral, remember where you heard it first. #TaylorSwift #Swiftie
0:35

Last year, almost to the day, Kayla Nicole went on Angel Reese podcast and talked about getting dumped by an NBA player and dressing up as the girl she got dumped for for Halloween. That was Teyana Taylor that she was shading. A week ago, Taylor Swift went on Jimmy Fallon and said the last movie she watched was the one starring the one and only, Teyana Taylor lol. She praised all the actors. Taylor swift is a GD mastermind and I love it. I have seen NO ONE catch this…. If it starts going viral, remember where you heard it first. #TaylorSwift #Swiftie

Nosey Ren

23,780 views • 9 months ago

GPT-5.5 is MUCH more reliable on longer running tasks - for the first time with any model. As we speak I have a migration running for over 7+ hours - this literally never happened before, the models would maybe run for 30 mins or of you really shout at them for 2-3 hours. Last night I went to sleep, set a long running task, then queued up 10 prompts to 'keep it going'. It did not stop after the first prompt and kept going for 8+ hours and I woke up to all the same prompts still queued up. The ability to run for a long time, in combination with ability to validate with computer use & other tools, makes it much more useful for building real applications.
0:15

GPT-5.5 is MUCH more reliable on longer running tasks - for the first time with any model. As we speak I have a migration running for over 7+ hours - this literally never happened before, the models would maybe run for 30 mins or of you really shout at them for 2-3 hours. Last night I went to sleep, set a long running task, then queued up 10 prompts to 'keep it going'. It did not stop after the first prompt and kept going for 8+ hours and I woke up to all the same prompts still queued up. The ability to run for a long time, in combination with ability to validate with computer use & other tools, makes it much more useful for building real applications.

Peter Gostev

105,634 views • 2 months ago

HOLY MOLY running a 35B model locally on a MacBook shouldn’t be THIS FAST 🤯 Spent my weekend in atomic.chat testing Qwen 35B vs. Qwen 27B on my local machine. I had them generate a fully animated HTML/Canvas car mini-game (demo below), ... and both models breezed through the physics and parallax scrolling without a hitch! The secret sauce here is the Atomic Chat app. Because it's perfectly optimized for Mac and uses Google's new TurboQuant under the hood, you can run heavy open-source models flawlessly while keeping top-tier output quality 👊 Other perks: → ZERO setup required → Access 1,000+ models completely free → 100% offline and private → Zero API limits ... and MUCH more! I dropped the prompt I used in the 🧵↓ Spin it up locally and let me know what you get!
0:11

HOLY MOLY running a 35B model locally on a MacBook shouldn’t be THIS FAST 🤯 Spent my weekend in atomic.chat testing Qwen 35B vs. Qwen 27B on my local machine. I had them generate a fully animated HTML/Canvas car mini-game (demo below), ... and both models breezed through the physics and parallax scrolling without a hitch! The secret sauce here is the Atomic Chat app. Because it's perfectly optimized for Mac and uses Google's new TurboQuant under the hood, you can run heavy open-source models flawlessly while keeping top-tier output quality 👊 Other perks: → ZERO setup required → Access 1,000+ models completely free → 100% offline and private → Zero API limits ... and MUCH more! I dropped the prompt I used in the 🧵↓ Spin it up locally and let me know what you get!

Charly Wargnier

100,007 views • 2 months ago

As for my summit dreams, they will have to wait There are no guarantees in mountaineering - I made the call at the South Col with respect to the Everest summit, that I was not in a sufficiently fit state to make an attempt, so I did not head off for the summit push. Given the dangers involved, I could not in good conscious make an attempt on the summit and potentially put myself and my Sherpa in peril. The mountain takes a swift toll on people who fail to leave sufficient energy in reserve for the descent. I tried to remain positive and head for the Lhotse summit on 24 May 2023 - the mind was willing but my body was not on the same page unfortunately, so I made a retreat for the safety of camp 2 2/3
0:13

As for my summit dreams, they will have to wait There are no guarantees in mountaineering - I made the call at the South Col with respect to the Everest summit, that I was not in a sufficiently fit state to make an attempt, so I did not head off for the summit push. Given the dangers involved, I could not in good conscious make an attempt on the summit and potentially put myself and my Sherpa in peril. The mountain takes a swift toll on people who fail to leave sufficient energy in reserve for the descent. I tried to remain positive and head for the Lhotse summit on 24 May 2023 - the mind was willing but my body was not on the same page unfortunately, so I made a retreat for the safety of camp 2 2/3

Trent Thorne

20,258 views • 3 years ago

Thalians, our girl needs us now more than ever🙏🏼Remember this will be the last time YOU & I VOTE for her, let’s triple our efforts so when Sunday comes, we can pat ourselves on the back & say we did that!🥹 • It’s always 10 numbers per device • ⁠Change to desktop site after voting to get extra 100votes. 1number =200votes • Once you’ve used 10 numbers on one device, DO NOT vote again on that device • If you still want to vote, switch to a different device and use new numbers 🔗: #ThandekaTshabalala #BBMzansi #BBMzansiS6
0:15

Thalians, our girl needs us now more than ever🙏🏼Remember this will be the last time YOU & I VOTE for her, let’s triple our efforts so when Sunday comes, we can pat ourselves on the back & say we did that!🥹 • It’s always 10 numbers per device • ⁠Change to desktop site after voting to get extra 100votes. 1number =200votes • Once you’ve used 10 numbers on one device, DO NOT vote again on that device • If you still want to vote, switch to a different device and use new numbers 🔗: #ThandekaTshabalala #BBMzansi #BBMzansiS6

Thandeka Tshabalala

35,440 views • 4 months ago

We are bootstrapping our app studio to $1M/mo It’s based around screen time apps. We have prayer lock with 13k reviews ranking top #100 on the AppStore Now we are building step lock our 2nd app and the first goal is to scale it. to $10k/mo, In public It combines health kit + screen time api To block your apps until you get your steps in. We used Rork plans feature, and it literally gave us the whole setup for the app, Then it built the app for us and one shorted the api integrations In a 2 days we had the app submitted to the AppStore The app is currently in review. Apple has rejected it 7 times For the dumbest reasons. But it’s looking like this time it will get approved. Will be posting updates on this
0:12

We are bootstrapping our app studio to $1M/mo It’s based around screen time apps. We have prayer lock with 13k reviews ranking top #100 on the AppStore Now we are building step lock our 2nd app and the first goal is to scale it. to $10k/mo, In public It combines health kit + screen time api To block your apps until you get your steps in. We used Rork plans feature, and it literally gave us the whole setup for the app, Then it built the app for us and one shorted the api integrations In a 2 days we had the app submitted to the AppStore The app is currently in review. Apple has rejected it 7 times For the dumbest reasons. But it’s looking like this time it will get approved. Will be posting updates on this

Ernesto Lopez

62,814 views • 4 months ago

I wish to sincerely apologise to the residents of Lakeside and its environs for the inconvenience caused. There have been multiple faults on cables connected to the feeder, one of which was located in the middle of a road. As a result, excavation works took some time to safely access and isolate the fault. Repair works are currently underway. Our technical team remains on site and will continue working until power supply is fully restored.
0:14

I wish to sincerely apologise to the residents of Lakeside and its environs for the inconvenience caused. There have been multiple faults on cables connected to the feeder, one of which was located in the middle of a road. As a result, excavation works took some time to safely access and isolate the fault. Repair works are currently underway. Our technical team remains on site and will continue working until power supply is fully restored.

Richmond Rockson, Esq.

45,162 views • 4 months ago

Cancelled ChatGPT -> Built JARVIS -> Pays $0 -> it works offline + it's smarter than the $20/month version. No WiFi needed, no cloud, no API keys, no rate limits, no queues, no $20/month just to ask a server in Virginia for the weather. Just a local model running directly on the laptop hardware, voice activated, system integrated, controlling apps, answering questions, doing the work. Iron Man had JARVIS embedded in his suit, this guy has it embedded in his MacBook and it works on a plane, in a basement, on a remote cabin with zero signal. OpenAI is burning $700,000 a day on infrastructure to deliver something this guy runs for free. Anthropic charges $200/month for unlimited Claude access, microsoft built Copilot into every product they sell. This guy skipped all of it, downloaded a model and made his laptop the smartest device in the room. No subscription. No login. No internet. No data sent anywhere ever. The most powerful AI assistant on earth is now the one running locally on hardware you already own. ChatGPT charges you to think slower, he pays nothing and thinks alone, he made it himself.
0:15

Cancelled ChatGPT -> Built JARVIS -> Pays $0 -> it works offline + it's smarter than the $20/month version. No WiFi needed, no cloud, no API keys, no rate limits, no queues, no $20/month just to ask a server in Virginia for the weather. Just a local model running directly on the laptop hardware, voice activated, system integrated, controlling apps, answering questions, doing the work. Iron Man had JARVIS embedded in his suit, this guy has it embedded in his MacBook and it works on a plane, in a basement, on a remote cabin with zero signal. OpenAI is burning $700,000 a day on infrastructure to deliver something this guy runs for free. Anthropic charges $200/month for unlimited Claude access, microsoft built Copilot into every product they sell. This guy skipped all of it, downloaded a model and made his laptop the smartest device in the room. No subscription. No login. No internet. No data sent anywhere ever. The most powerful AI assistant on earth is now the one running locally on hardware you already own. ChatGPT charges you to think slower, he pays nothing and thinks alone, he made it himself.

Defileo🔮

154,009 views • 2 months ago

Klay Thompson on his exchange with Ja Morant: “Nothing of intelligent depth. It was really just running his mouth, and he’s been running his mouth for a long time, and it’s funny to run your mouth when you’re on the bench. It’s kinda the story of his career so far… He’s just been letting a lot of other stuff get in the way of that…When you’re a star, it comes with great responsibility, and I hate to see that go to waste.” (via Joey Mistretta)
0:15

Klay Thompson on his exchange with Ja Morant: “Nothing of intelligent depth. It was really just running his mouth, and he’s been running his mouth for a long time, and it’s funny to run your mouth when you’re on the bench. It’s kinda the story of his career so far… He’s just been letting a lot of other stuff get in the way of that…When you’re a star, it comes with great responsibility, and I hate to see that go to waste.” (via Joey Mistretta)

Legion Hoops

7,737,123 views • 7 months ago

.Fiat and I go wayyyyy back. Their cars are featured in the music videos for This Kiss, Good Time, Cut to the Feeling and now On Wires: Fun fact: Fiat is the very first car I owned when I first moved to LA. For ten years I ripped around in that bad boy. I wasn’t what one would call “car savvy” and so it was about three years in, during my drive to a session one day that I finally tried the button for the sun roof. I was on the phone w my dad at the time, middle of the highway and started screaming…”oh my gawd I have a convertible!” That story lives on as one of my family’s fav “oh Carly” moments. Anyways, thanks to Fiatusa for all the support over the years. And for that hot little convertible x
0:15

.Fiat and I go wayyyyy back. Their cars are featured in the music videos for This Kiss, Good Time, Cut to the Feeling and now On Wires: Fun fact: Fiat is the very first car I owned when I first moved to LA. For ten years I ripped around in that bad boy. I wasn’t what one would call “car savvy” and so it was about three years in, during my drive to a session one day that I finally tried the button for the sun roof. I was on the phone w my dad at the time, middle of the highway and started screaming…”oh my gawd I have a convertible!” That story lives on as one of my family’s fav “oh Carly” moments. Anyways, thanks to Fiatusa for all the support over the years. And for that hot little convertible x

Carly Rae Jepsen

268,986 views • 12 days ago

Huge announcement. I'm leaving Bloomberg as an OCaml Engineer. And.. I'm joining Bloomberg as a C++ Engineer. After using Functional Programming primarily for the last 10 years, it's time for me to switch gears. I moved to another team internally. I'll be working on low-latency real-time market data in C++. It's a big change for me. But I'm excited about a new challenge! AMA
0:11

Huge announcement. I'm leaving Bloomberg as an OCaml Engineer. And.. I'm joining Bloomberg as a C++ Engineer. After using Functional Programming primarily for the last 10 years, it's time for me to switch gears. I moved to another team internally. I'll be working on low-latency real-time market data in C++. It's a big change for me. But I'm excited about a new challenge! AMA

Dmitrii Kovanikov

301,280 views • 1 year ago

Streaming iPhone data in real-time directly to Rerun 🚀 The collection process is one of the most frustrating parts of building imitation-learning datasets. I’ve got a little army of sensors—📱 iPhone, iPad, Quest 3—but getting them temporally aligned, spatially aligned, AND seeing real-time feedback while recording is tough. I stumbled on a great library from Cake Lab (WPI) called ARFlow. It’s a thin client built on Unity’s ARFoundation that connects over gRPC to a server running Rerun for live data logging. I forked it to: - Log the SLAM translation poses, and - Upgrade rerun to v0.23 for my use case. So far, it works well, but there are still a few hitches: 1. Right now, it’s solid on iPhone and iPad; my Quest 3 client is still slow and not super reliable. 2. I’m using an older ARFlow branch focused on real-time streaming only—no spatial or temporal sync yet. Unity builds for iOS keep failing. 🛠️ 3. Nothing is saved locally to the client, so packet loss is a risk on shaky networks. There’s huge potential in tapping the ubiquitous sensors we carry around every day, and ARFlow is a big step toward making that easy
0:16

Streaming iPhone data in real-time directly to Rerun 🚀 The collection process is one of the most frustrating parts of building imitation-learning datasets. I’ve got a little army of sensors—📱 iPhone, iPad, Quest 3—but getting them temporally aligned, spatially aligned, AND seeing real-time feedback while recording is tough. I stumbled on a great library from Cake Lab (WPI) called ARFlow. It’s a thin client built on Unity’s ARFoundation that connects over gRPC to a server running Rerun for live data logging. I forked it to: - Log the SLAM translation poses, and - Upgrade rerun to v0.23 for my use case. So far, it works well, but there are still a few hitches: 1. Right now, it’s solid on iPhone and iPad; my Quest 3 client is still slow and not super reliable. 2. I’m using an older ARFlow branch focused on real-time streaming only—no spatial or temporal sync yet. Unity builds for iOS keep failing. 🛠️ 3. Nothing is saved locally to the client, so packet loss is a risk on shaky networks. There’s huge potential in tapping the ubiquitous sensors we carry around every day, and ARFlow is a big step toward making that easy

Pablo Vela

27,383 views • 1 year ago

i'm too stubborn to simply give up on the concept of Benji after almost 10 years(!!) of iterating on it with different stacks... 2018: graphql backend + react SPA 2019: made a react native app for 6 months 2020: gave up on it and wrote 3 years of using other bs tools and being grumpy about them 2023: rewrote it in blitz js (JUGE mistake) 2024: iterate on react native + PWA + RN pwa wrapper 2025: time for a rewrite to Zero To Shipped ⛴️ stack (but next js sucks for this kind of app, add api, mcp, cross platform app, and many other things 2026: i want it all actions INSTANT AND FAST AF and next js is the wrong tech choice for a mostly client app, so i'm starting a rewrite with TANSTACK + Convex + Better Auth it feels so snappy 😭 btw 99% of the customers abandoned it, there are like 3 ride or die fans but i know i can turn it around!!!! let /goal cook.
0:18

i'm too stubborn to simply give up on the concept of Benji after almost 10 years(!!) of iterating on it with different stacks... 2018: graphql backend + react SPA 2019: made a react native app for 6 months 2020: gave up on it and wrote 3 years of using other bs tools and being grumpy about them 2023: rewrote it in blitz js (JUGE mistake) 2024: iterate on react native + PWA + RN pwa wrapper 2025: time for a rewrite to Zero To Shipped ⛴️ stack (but next js sucks for this kind of app, add api, mcp, cross platform app, and many other things 2026: i want it all actions INSTANT AND FAST AF and next js is the wrong tech choice for a mostly client app, so i'm starting a rewrite with TANSTACK + Convex + Better Auth it feels so snappy 😭 btw 99% of the customers abandoned it, there are like 3 ride or die fans but i know i can turn it around!!!! let /goal cook.

kitze · supermac.io 🐦‍🔥

16,580 views • 1 month ago

Introducing Replit ModelFarm, the fastest and safest way to build your next Generative AI app. Available for free on Hacker and Pro plans till October 15th. It requires zero setup, zero configuration, and zero API keys. With Replit ModelFarm, you can build a working Gen AI app in as little as 3 lines of code. Get started by installing the Replit AI library in any Python, JavaScript, or TypeScript Repl. The library implements an API for text completion, chat completion, and text embeddings. It supports streaming so your users can see model responses in real-time rather than waiting on a single output. All Hacker and Pro builders will have free access to a selection of Gen AI models offered by Google Cloud Vertex AI through Replit ModelFarm. All models are accessible from the development environment and any deployed app.
0:16

Introducing Replit ModelFarm, the fastest and safest way to build your next Generative AI app. Available for free on Hacker and Pro plans till October 15th. It requires zero setup, zero configuration, and zero API keys. With Replit ModelFarm, you can build a working Gen AI app in as little as 3 lines of code. Get started by installing the Replit AI library in any Python, JavaScript, or TypeScript Repl. The library implements an API for text completion, chat completion, and text embeddings. It supports streaming so your users can see model responses in real-time rather than waiting on a single output. All Hacker and Pro builders will have free access to a selection of Gen AI models offered by Google Cloud Vertex AI through Replit ModelFarm. All models are accessible from the development environment and any deployed app.

Replit ⠕

229,285 views • 2 years ago