JD Prater

Introducing Universal Streaming - an ultra-fast, ultra-accurate streaming speech-to-text model for voice agents 🚀 Universal-Streaming delivers ultra-low latency, superior accuracy, and intelligent endpointing at just $0.15/hr 👇

🚀 Big news from AssemblyAI! We’re unveiling two major product breakthroughs set to redefine Speech AI capabilities in 2025. Speech AI adoption is skyrocketing, and we’re here to fuel that growth. Here’s what’s coming: 🧵👇

AssemblyAI's founder Dylan Fox was featured in @googlecloud's recent AI Trends report, discussing the real challenges of bridging AI's "last mile" gap from demo to production 🧵 #FutureofAIStartups2025

AssemblyAI is featured in a @googlecloud case study! Learn how we build & train our Speech AI models faster by leveraging Google Cloud and Vertex AI: cloud.google.com/customers/asse…

@chiefaioffice Yes please

Maintaining a database of VC rounds in AI this year 1500+ rounds, $55B+ raised Comment if you'd like access

The Universal-2 model was just released. This is a next-generation Speech-To-Text model. It's a 660M parameter Conformer RNN-T model built on Universal-1's (already great) performance. Some results: • It's 24% better at recognizing proper nouns • It's 21% improvement in alphanumeric accuracy • 15% enhanced text formatting • 73% of users prefer Universal-2 compared to Universal-1 • More accurate and robust, especially in real-world speech complexity This model solves some of the hardest challenges in making Speech-To-Test reliable For the first time, we can say that WER metrics (Word Error Rate) don't capture the full story about how good this model is. You can go here to try this model: assemblyai.com/playground And here is a blog post with their announcement: assemblyai.com/universal-2 @AssemblyAI is behind Universal-2. I am a long-time collaborator of the company.

Announcing Universal-2, our most advanced speech-to-text model to date. Building on Universal-1's industry-leading performance, we've pushed Speech AI even further. Here's what's new:

@arvidkahl @lukemckaysmith Have you tried @AssemblyAI? I’d love to know your thoughts? assemblyai.com/benchmarks

Not at all. My moat is not how fast I transcribe, but how to extract actionable data from it (and how to build an infrastructure that can handle it.) If I had a few million, I'd just buy more compute to speed things up. The hidden knowledge that took me a year to learn is what makes this hard to copy.

OpenAI just released a new transcription model called "whisper-large-v3-turbo" — which is 8 times as fast(!!!) as their most recent highest-quality model. 🤯 I am currently deploying this to my fleet of servers. This single change will likely 2x-4x the number of podcasts I can handle per day. Besides that, it will increase the baseline quality of ALL transcriptions on Podscan. It will EVEN allow me to go through my backlog much faster. That means faster alerts and more historical search data for Podscan.fm customers. OpenAI is the gift that keeps on giving.

Introducing Assembly Required - A series of candid conversations with AI founders sharing insights, learnings, and the highs and lows of building a company. This week, @EdoLiberty, founder and CEO of @pinecone, sits down with @YouveGotFox, founder and CEO of AssemblyAI, to talk through: - Pioneering an entirely new category that doesn’t even have a name yet - Making the right pricing decisions, even when they’re hard - Building products that are optimized for love and trust Watch their full conversation here: assemblyai.com/assembly-requi…

Such an insightful conversation.

@torylynne @thetafferboy @samtorresatl I did! Leading product marketing at @AssemblyAI working on speech-to-text.

Question: I have 100M rows of query data, each query has been assigned a category. If I wanted to build a classifier trained on this data, that means you could enter a query and it would classify it into one of those categories, based on that data - how would you go about it?

@Ameobea10 Or from the mind of Tim Burton

Saw this awesome statue while out walking today Amazing vibes- looks like something out of a video game or surreal art piece and was very neat to see in the real world

@bentossell Veed and Descript

best video editing tool? need: - auto-detect speakers - transcription, timestamps + speakers - edit video from transcript - simple way to layer on a cropped version of the video (overlay the speaker when presenting screen) - show notes nothing seems to do all this well + simply

Let’s go!

@DigitalSamIAm What’s with the hand gestures?

X: Everyone - people, companies, organizations - deserves the right to have free speech. X: Owner boards off the crazy train + burns bridges with numerous core constituencies of major companies. Everyone: OK, we're going to exercise our free speech and not advertise on a platform whose beliefs don't align with our own. X: We didn't mean free speech like that. That's not good free speech. We're going to sue to make them do the good free speech on X.

Looking for an easy no-code option to transcribe audio files? We're excited to launch our @zapier integration v. 2.0! The enhanced app is more flexible, supports more Speech AI features, and integrates more closely into the Zap editor. assemblyai.com/blog/updated-z…

@nickfloats @midjourney @krea_ai @topazlabs @elevenlabs @suno_ai_ @LumaLabsAI @splinetool @descript @MoveAI_ For speech, use @AssemblyAI

For images, use @midjourney For real time, use @krea_ai For upscaling, use @topazlabs For voice, use @elevenlabs For music, use @suno_ai_ For NeRFs, use @LumaLabsAI For 3D, use @splinetool For editing, use @descript For mocap, use @MoveAI_ For video, use prayers

🤖🔧 In this video, we take a look at the best AI tools for software engineers. Specifically for coding and some supporting tasks like testing and UI building. These tools are used by working professionals in many fields such as web development, python and java development, game development, app development and much more. ⭐️ Watch: youtu.be/66rQhuMV1fg

Don't underestimate yourself if you: -take notes with action items & share with the group -review, make comments, resolve comments to move things forward -write outline for plan of attack, share it -proactively send status updates in slack you are already ahead of 90% of people.