Natural Language Processing
topic on 12 shows · 31 statements across 29 episodes
In Depth 
American Optimist 
Latent Space 
My First Million 
No Priors 
the Official SaaStr Podcast 
A Product Market Fit Show 
Capital Allocators 
Top Founders 
the MAD Podcast 
the a16z Podcast 
20VC
31 statements about Natural Language Processing, every show
Socher: Majority of his early 2010 neural network NLP papers were rejected
“The way careers work is you want to be kind of novel, but if you're too novel, if you're too far out there, then your papers will get rejected, and that certainly happened to me a lot in the early days, like, 2010 of neural networks for natural language proces…”
Jain: CSFP Traded Off Natural Language Processing In Late 1990s
“We were doing natural language processing back in the late nineties. We were taking in the news feeds and trading things off of it.”
Madheswaran: BERT cut entity extraction training data requirements to ten documents
“So, right, so you went from needing to train thousands about thousands of documents to 10 documents, maybe, to extract things like dollar amounts from invoice totals simple things like that.”
Drew Houston: Regular expressions solve 95% of email triage NLP tasks
“I tried to do all these like NLP things and then to my dismay, like a bunch of regexes for like, got you like 95% of the right there.”
Delangue: Non-text AI modalities will transform technology like NLP did
“It's gonna be audio, it's gonna be video, it's gonna be biology, chemistry, AI. These are, in my opinion the domains that in the next few years are going to change the world the same way NLP has in the past.”
Hunt: Core Math Behind Narrative NLP Has Not Changed In 35 Years
“Nothing's changed in the math. It's the same algorithms, it's the same network math that I was using 35 years ago. What has changed is big data, so the access to all this information, all this text information.”
Liu: Admitted being bearish on LLMs for four years before recent breakthroughs
“I mean, the biggest one really was the fact that, like, I think for just four years I was so bearish on language models. And just NLP in general, I was just like, ah, like, none of this really works. Like, why would I spend time focusing on this? I gotta go do…”
Sapoznik: OpenAI is aggressively commoditizing NLP and machine learning capabilities
“Where we see this going is what open AI and all that we're doing is really a really fantastic and spectacularly aggressive commoditization of NLP and some other machine learning capabilities.”
Masad: 2012 'Naturalness of Software' paper proved NLP models apply to code
“There was this seminal paper in 2012 called on the naturalness of software. And basically a bunch of researchers try to apply NLP to code. And what they found is that actually code Can be modeled like any language. That's why they call it naturalness is becaus…”
Machine learning engineering skills transfer broadly across vision, NLP, and audio domains
“One of the things about machine learning and AI in this industry is that there's a lot of crossover between these different data domains. You know, if you're able to solve a computer vision problem pretty well, a lot of those skills are transferable to natural…”
The only meaningful way to adopt NLP and LLMs is cloud-based
“The only meaningful way to adopt to adopt NLP and LLMs is in the cloud”
Socher: AI advances faster by expanding unified models than building separate ones
“Kind of like imagine Wikipedia and everyone just keeps adding to Wikipedia and keeps making one dictionary better rather than everyone who wants to build a dictionary just starts their own dictionary company and then builds it from scratch. It just doesn't mak…”
Gil: Generative AI is a technology disruption, not an extension of prior ML
“And in reality, we've had a technology disruption. We've shifted to two very different architectures, diffusion-based models, which is a statistical physics model for ImageGen. And then on the language side, we moved to these large language models, which some …”
Delangue: NLP, vision, and audio are the top three Hugging Face tasks
“The three main tasks right now are NLP, so text, right?
From like information extraction, text generation, text classification.
the second one is text to image and computer vision, right?
So object detection, text to image, text image generation.
the third o…”
Liang: Foundation Models Cause the Concept of an AI Task to Dissolve
“And this was a paradigm shift, in my opinion, because it changed the way that we conceptualize Machine learning and NLP systems from these bespoke systems where you're, it's trained to do question answering, to train to do this, to just a general substrate whe…”
Pande: Insurers will use NLP to process AI-generated reimbursement letters
“Eventually someone has to read all those letters. Someone has to validate them, and it's probably, you know, some sort of NLP on the other side.”
Momand: Esper built an NLP database of US and Canadian regulations
“So basically when we started us where we invested a ton of time into building up this like database and ontology of regulations across the federal state and Canada, and we've applied some cool NLP and other, you know, tech things to analyze the data and find s…”
Shah predicts natural language processing will be a bigger mega-trend than mobile
“And all we really need is a translation layer, and so this is a broad-based, in my mind a mega trend to be, which is, In my mind, bigger than mobile.”
Delangue: NLP is the most important field of machine learning
“I believe that NLP today is the most important field of machine learning.”
Delangue: AI is ending human exclusivity over natural language processing
“We're really moving from a world where only humans could do natural language, which is why you have search large customer support, sales, community, communication departments in companies.”
Delangue: Every company will use NLP within three years or face disadvantage
“I believe that in in three years, every single company in the world will use NLP. And even more than that, I believe that in three years, a company that is not using NLP will be in a systematic disadvantage compared to a company that is using NLP, because it m…”
Turakhia: Clinical documentation terminology varies between individual hospitals
“In natural language processing that's embedded in AI, the lexicon that people use, how doctors and clinicians write what it is that they're seeing with their patient is different from not even specialty to specialty, but hospital to hospital, sort of mini subc…”
Sinofsky: State-of-the-art machine translation still relies on 1970s NLP
“The best example for me of that is how everybody said machine learning was going to replace all of natural language processing. But if you dig into any of the work that's been going on, even the most state of the art translation, which, you know, goes any lang…”
Thakur avoids buzzwords like NLP and machine learning in system design
“Notice I did not quite use buzzwords like NLP, or deep learning, or data science, or machine learning, which I could very well have used, but I don't think it's that useful.”
Davis: NLP tasks hard for humans remain hard for AI
“Basically anything that is really hard for you to do as a human is still gonna be pretty hard for us.”
Tunguz: Off-the-shelf AI models only hit 80% of required enterprise performance
“You can take a generic NLP or speech learning model and you can get 80% of the way there, but in order to really have a fundamentally new experience, you need to get it 95% of the way there, and you can't do that with just off-the-shelf algorithms. You're goin…”
Snip.ly is building an AI tool to turn blogs into videos
“Now we're investing very heavily in a new platform. That's going to automatically convert blog posts into videos using natural language processing and a bunch of artificial intelligence.”
MetaMind developed a single deep learning model for multiple NLP tasks
“The really amazing thing and the reason why I'm saying there's a bit of a new breakthrough coming up now is that we can actually solve a lot of different NLP problems now with the same kind of deep learning model, and this is one that we've just developed A co…”
Qlik plans to integrate natural language processing without building underlying models
“So, natural language processing the simplest answer is yes. The longer answer is processing the language isn't the hard, I mean, that's the hard part. It's not one we're gonna do. Dealing with the language once you have it is something we're obviously very int…”
LeCun predicted in 2014 deep learning would conquer natural language processing
“There's a kind of a sense that in the community that the next set of techniques to kind of fall to deep learning, if you want, will be natural language processing.”