Everything Maren Nelson said on any show that made the record, most notable first. Each card names its show and opens the statement there.
Nelson: Human contractors will remain in Clara's loop until AGI exists
“I think they will always be involved. You know, it's, we get asked that frequently, and it makes perfect sense. For us, though, until we have some kind of strong AI that completely obfuscates the need for human work, you necessarily have humans in the loop.”
Nelson: Grouping all AI products together is as flawed as web apps
“I think the problem is largely with thinking about those things as synonymous to each other. It's kind of like saying that everybody who builds a web app is competing in the same category, right? It's kind of like saying that just a software app like MyFitness…”
Nelson: AI-native startups possess superior ML environments over legacy software firms
“I do think that there is an advantage to many of the younger companies that have started from their foundation with AI at the center of everything that it is that they're doing, because they have created much better environments to support machine learning the…”
Nelson: Pushing ML past 80-90% accuracy is the primary product challenge
“It's letting machine learning get past the kind of stuck in 80 to 90% right land that isn't good enough to be shipped as a user facing feature. Getting over that hump is, is a huge challenge.”
Nelson: Tech adoption is far more incremental than people predict
“I think everything tends to be much more incremental than we like to think of it before it happens.”
Nelson: Snapchat succeeds in AR by grounding innovation in familiar behavior
“Snapchat has this Relentless product focus. And now, of course, they're the people introducing the idea of augmented reality to tons and tons of people, and games like Pokemon Go, right, where there is this slow but sure corralling of technology-enabled innova…”
Nelson: Language interfaces require high task complexity to outperform GUIs
“Language interfaces are really important if there is a sufficient amount of complexity such that language is the best interface to resolve that complexity.”
Nelson: Simple chatbots provide worse UX than app buttons like Uber
“And unfortunately, bots today are not sophisticated. They don't understand nuance or complexity. And so the application for these bots tends to be things that are really, really simple. But of course, it's easier for me to click the Uber app, right, and get an…”
Nelson: Machine learning excels at repetitive, low-variance NLP tasks like scheduling
“So scheduling, luckily, is very much one of them. Repetitive, right? You want the variance in response to be low, ideally. So in our case, you know, you're talking about the same handful of variables on every occasion. You have the ability to kind of constrain…”
20VC 1