stephensonsco

6 ms·

1.
The history of the word “hacker”(blog.deepgram.com)
2 points by stephensonsco 8y ago|0 comments
2.
by stephensonsco 8y ago
I didn't answer "Do you have a cheap way of generating high quality data?". We have good ways to do it. They're not that cheap though. It's expensive (organizationally and real $$$) to label large amounts of data no
3.
by stephensonsco 8y ago
Yep, all good points. One thing to consider is that generalization is a big problem. It's easy to get good on a specific dataset nowadays (like 5-10% word error rate level on academic datasets), but that same model might do 40% WER on
4.
by stephensonsco 8y ago
You can pay to get training at a lower usage amount too.
5.
by stephensonsco 8y ago
You find out very interesting things even randomly sampling your life in audio. We still come back to this for fun. The original device was an intel edison but recent variants have been based on the raspberry pi zero w.
6.
by stephensonsco 8y ago
If you're training from scratch around 10k hours is needed to get a good model, but when you are transfer learning you don't need nearly that much (100 hours gets you a lot). We excel in phone call and meetings settings. I.e. the
7.
by stephensonsco 8y ago
There's no additional charge for training a custom model when your usage is a minimum of 10k hrs/mo.
8.
by stephensonsco 8y ago
Best to say "yes! but only some of the time". It's something we're working on right now. You can be 80% accurate, by some metric, but it's still not good enough usually to pass a human's sniff test. Good speake
9.
by stephensonsco 8y ago
We do custom models (train the full DNN, not just tack on a new text language model) using transfer learning and it works for small numbers of examples too. Glad to hear you asking about fuzzy search. That's something we do (it's
10.
by stephensonsco 8y ago
It's a metric that's hard to nail down because there is so much parameter space that you are flattening into one number. Also it doesn't address the "I care about these five high value words (that are made up), can you r
11.
by stephensonsco 8y ago
This is a seriously fertile area where you get to "define the new interface". It's a big problem though, since few buyers know they want those things. Around 95% of customers come into it with "give me the transcripts&qu
12.
by stephensonsco 8y ago
Price starts at $1/hr billed in 1 second increments. Frequently we charge less than that, since the price is dropped with volume, and that's typically businesses have a steady amount running through them (a few thousands hours). M
13.
Launch HN: Deepgram (YC W16) – Scalable Speech API for Businesses
56 points by stephensonsco 8y ago|23 comments
14.
by stephensonsco 8y ago
Pretrained models are very nice to spin up a system and start using it. You need that because training the model is so hard. But, the pretrained models are by no means a good general model. They are trained with narrow params (like #/s
15.
by stephensonsco 8y ago
You're doing awesome (arduous) work. The text normalization is especially a total bear. I feel your pain. Limiting your text to one file is good in many ways because it allows you to scope down the amount of work needed to do a compari
16.
Azure is Down for A.I. Companies(scottiam.com)
1 points by stephensonsco 8y ago|0 comments
17.
by stephensonsco 8y ago
Anyone able to find any data on time 'til accuracy? I don't see it (even in the video linked in another comment). Sure it's nice to "achieve 90% scaling efficiency" in images/sec, but images/sec alone does
18.
by stephensonsco 8y ago
The dark matter is a gas not a solid. It's not orbiting the sun, it's orbiting the galaxy (technically, not even the galaxy, it's orbiting inside it's own halo mostly). The orbits of the particles would be highly irregul
19.
by stephensonsco 8y ago
Also very true. Unfortunately this data is dirty, context dependent, or plain missing. This is true of pretty much any "real" experiment.
20.
by stephensonsco 8y ago
This is awesome. But it is still impractical. The researchers knee deep in this already have to fight for the computing power and data access (and error checking of all of that) for years to get anything meaningful to come out. It's ju
21.
by stephensonsco 8y ago
In the direct-search-for-dark-matter-experimental-community the DAMA results have long been excluded and therefore ~discredited (see lots of papers from LUX/XENON/PANDAX/CDMS/etc). But theoretician's have to keep be
22.
by stephensonsco 8y ago
So you think the LHC should "publish" 100 petabytes of data? What you are stating isn't practical because of cost. But I'll definitely go with the idea that "if you want to make a claim as big as finding dark matter
23.
by stephensonsco 9y ago
The problem is that a founder who hasn't already filled that role themselves doesn't know two things: 1) what the key parts of doing the job are and 2) which skills/personality the hire must possess. There are many many ways
24.
by stephensonsco 9y ago
build it, you'll have a hit
25.
by stephensonsco 9y ago
Coolest A/V project I've seen in a while! You can bump up the playback speed in the youtube player to compress temporally even more. Definitely put in a "Click Here to Try A Sample Video" as others suggested.
26.
by stephensonsco 9y ago
Deepgram (YC W16) is hiring for frontend. DG trains and deploys deep neural speech networks to enterprise and public APIs with state-of-the-art spoken language analysis. We care a lot about building products that are fast, accurate, cutting
27.
by stephensonsco 9y ago
Yep, it is only one trick. Just like electrifying the world only had that one weird trick of alternating current transported miles with metal cables and the computing world that trick of transistors. Pretty good tricks though. Understanding
28.
by stephensonsco 9y ago
Deepgram (YC W16) is hiring for frontend and sales. DG trains and deploys deep neural speech networks to enterprise with state-of-the-art spoken language analysis. We care a lot about building products that are fast, accurate, cutting edge,
29.
by stephensonsco 9y ago
Nope, not a must have! Just bring in depth problem solving and architecting skills.
30.
by stephensonsco 9y ago
Deepgram (YC W16) is hiring for frontend, backend, A.I., and sales. DG trains and deploys deep neural speech networks to enterprise with state-of-the-art spoken language analysis. We care a lot about building products that are fast, accurat

More ›