Assertion Supported AI assessment confidence: 95% certainty 4/5 debate potential 1/5

Fox: AssemblyAI's upcoming speech model trained on 12.5M audio hours

Dylan Fox · This solo founder bet on AI 7 years ago. Now he has 5,000 customers & $115M raised. | Dylan, Foun... · PMF Show · Apr 8, 2024 · at 4:40

AssemblyAI founder Dylan Fox discusses the scaling of speech AI models with host Pablo Srugo, comparing his first model to AssemblyAI's latest model.

0:00 / 0:16exact quote · 16.3s
▶ Watch the full episode on YouTube → 720p mp4 · rendered on demand · StarZero watermark
“The first model I ever trained for speech to text Was on 10,000 hours of audio data. And then now the model that we're going to release soon, that's trained on 12 and a half million hours of audio data. It's on, you know, hundreds of TPUs.”

quote is from the automated transcript, cleaned for reading: filler sounds and stutters are removed, nothing is rephrased. names can be misheard (the analysis reads context, assessments check outside sources). how →

More from Dylan Fox

What-if
Fox: AssemblyAI would have moved faster by prioritizing accuracy over early GTM
“In the early days, I think we could have actually gone a lot faster if we had said, okay, let's do nothing but pick accuracy metrics that we feel like we want to surpass. And once we surpass those metrics, then let's go and Take this thing to market.”
Dylan Fox Apr 8, 2024 ▶ 19:48 This solo founder bet on AI 7 years ago. Now he has 5,000 customers & $115M raised. | Dylan, Foun... · PMF Show
Insight
Fox: Beyond a certain point, more resources hinder product development
“I believed at the time and still believe that to create the best product is not just a function of resources. You know, you don't just like put a budget and people in a bottle and shake, and then out comes this like amazing product. You can definitely increase…”
Dylan Fox Apr 8, 2024 ▶ 10:55 This solo founder bet on AI 7 years ago. Now he has 5,000 customers & $115M raised. | Dylan, Foun... · PMF Show
Insight
Fox: Founders should privately draft answers to every anticipated investor objection
“Every time I fundraise, and this is actually maybe a good tip for like startup founders that are fundraising, I actually write up like a big document with all of the objections I anticipate to get. And I write out like lengthy answers to all of them. And I don…”
Dylan Fox Apr 8, 2024 ▶ 15:50 This solo founder bet on AI 7 years ago. Now he has 5,000 customers & $115M raised. | Dylan, Foun... · PMF Show
Insight
Fox: RLHF and dialogue crossed the capability threshold that triggered LLM adoption
“It was the RLHF and the dialogue component of ChatGBT that triggered the takeoff. And that was the capability threshold. Like it was at that time that this capability threshold passed where now it is a prior, like LLMs are a priority for every organization.”
Dylan Fox Apr 8, 2024 ▶ 19:07 This solo founder bet on AI 7 years ago. Now he has 5,000 customers & $115M raised. | Dylan, Foun... · PMF Show
Insight
Fox: First MVP customers can derail startups with disproportionate influence
“You might have a customer that will want to use your early kind of MVP, but if that customer pulls you in the wrong direction, they're going to have a ton of influence because they're your first customer. And then you have to build for them.”
Dylan Fox Apr 8, 2024 ▶ 22:18 This solo founder bet on AI 7 years ago. Now he has 5,000 customers & $115M raised. | Dylan, Foun... · PMF Show
Insight
Fox: Level 5 autonomous vehicles have guaranteed PMF, bounded only by unit economics
“Similar to self-driving cars, like, you know, that level five self-driving cars will have product market fit, even though they don't exist yet. Now, of course, if they cost a million dollars per vehicle, the addressable market's probably much smaller versus if…”
Dylan Fox Apr 8, 2024 ▶ 28:38 This solo founder bet on AI 7 years ago. Now he has 5,000 customers & $115M raised. | Dylan, Foun... · PMF Show
Made with StarZero

Turn any episode into a week of clips.

This entire site, over 250 episodes transcribed, diarized, checked and made playable, runs on the StarZero media pipeline. Drop in your own episode and the podcast clipper finds the moments worth sharing, cuts them, captions them, and reframes them for every feed.