Pinned Tweet
Argmax Pro SDK is now generally available!!
The biggest favor @ChatGPT could have done to @monogram_ai 🙏
Atila retweeted
We are hiring research interns who want to push the frontier of voice AI!
Check out our latest research here:
argmaxinc.com/blog/interspee…
Sydney is gorgeous.
Glad Interspeech 2026 was held there this year (this week). I was surprised that I had the following encounter with several Aussie shop owners:
Them: How did you hear about us?
Me: AI
Them: Who is that?
Me: ChatGPT
Them: Anyways, thanks for coming!
Atila retweeted
Introducing fal Recast, an easy way to change any character in one click.
Upload an existing video + a reference photo of your new character. fal Recast changes who’s on screen while preserving the original motion, camera movements, cuts, and audio. Create character variations, adapt campaigns for different audiences, all without another shoot.
Powered by H3 Max, the #1 model for overall quality, prompt understanding, and aesthetics against leading video models.
A big thank you to everyone who has been beta testing @Narraloud so far!
We just published v1.1 introducing Plus Voices on top of Standard Voices:
> Characters stay the same
> Contextual pronunciation improves drastically
> Emotional range is higher and more natural
Please give it a spin and submit your feedback form! If you are not yet in our beta program, DM me for access.
We have been building an iOS app in stealth and we just hit private beta today!
> Life-like voices running on your iPhone
> Listen to anything: Books, PDFs, articles, emails...
> Works offline as you commute or travel
Beta testers who submit feedback will get a free lifetime license. DM me if you want an invite.
I wish my first internship were this productive!
Here is our second @argmax intern who killed it this summer:
Our interns @argmax built a groundbreaking harness for open-weight TTS models!
Check out @RmdW_W's explainer thread below:
Open-weight TTS models can look remarkably close to the best closed-weight models on benchmarks and demos. But are they reliable enough to actually replace them?
We stress-test Qwen3-TTS and VoxCPM2 in real-world scenarios, uncover reliability failures that standard evaluations miss, and develop an inference harness that eliminates the reliability gap. 🧵
Things that are left ambiguous after yesterday's Apple keynote:
1) Does float8 support come with 2x throughput vs float16?
int8 support did not bring 2x throughput on ANE until A17 Pro. Switching from float16 to int8 only realized memory and energy savings in A14, A15 and A16. This also coincided with the ANE being a new and "maturing" piece of silicon and so float8 may still come with 2x throughput support in its first year.
2) What is the programming abstraction for multi-ANE?
The only precedent for this is the M1/2/3 Ultra chips, which have 2 ANEs. Ultra chips are architecturally different because they fuse 2 separate dies with 1 ANE each, whereas A20 Pro brings 2 ANEs onto the same die. Same-die placement makes it easier to offer tensor parallelism to run larger models by letting the lower-level code infra (ANE Compiler, ANE Driver) shard and dispatch them across multiple ANEs. This is the first thing we will validate in our benchmarks when we get our A20 Pro devices next month!
In the worst case, there will be request/data-parallelism, just like Mac Studio and it will be a throughput improvement for running multiple instances of a given model or multiple independent models (unless Apple reserves 1 of the 2 ANEs for itself).
We are about to see whether my prediction holds with iPhone 18 and Apple doubles down on the Neural Engine in 3 minutes. 🍿
Apple GPUs are much more powerful with the new Neural Accelerators in A19/M5+. But this upgrade did not significantly improve performance per watt (think sustainability for long-running tasks) to be competitive with the Apple Neural Engine.
Given that 2025 was a dormant year for the ANE and a major upgrade for the GPU, we are looking forward to A20/M6 ANE!
The ultimate answer to "How does Hugging Face make money?"
Congrats to @julien_c @ClementDelangue and dozens of other HF folks we collaborated with over the years!
Exciting day for NVIDIA and @huggingface.
Open models strengthen safety and cybersecurity, accelerate innovation and diffusion, and enable sovereignty. They allow every developer, startup, university, industry and country to build with, customize and benefit from AI.
Thank you @ClementDelangue for coming to me.
NVIDIA is going to be a great home for Hugging Face, its community and the future of open models. 🤗
blogs.nvidia.com/blog/nvidia…
Getting questions about battery life
Test:
- Generated audio for The Odyssey in seconds
- Listened to it for 11 hours 54 minutes overnight
- iPhone Air battery dropped from 100% to 76%.
We have been building an iOS app in stealth and we just hit private beta today!
> Life-like voices running on your iPhone
> Listen to anything: Books, PDFs, articles, emails...
> Works offline as you commute or travel
Beta testers who submit feedback will get a free lifetime license. DM me if you want an invite.