Tomasz Tunguz - The Rarity Shibboleth
Tomasz TunguzVenture Capitalist If you were forwarded this newsletter, and you'd like to receive it in the future, subscribe here. The Rarity Shibboleth
Large language models are wonderful at ingesting large amounts of content & summarizing. Uploading an academic paper and I can pester it with an infinite list of questions & it will respond with equally infinite patience. In comparing the two most recent Microsoft earnings calls, Claude highlighted:
Excellent analysis. Benn Stancil described LLMs as great averagers of information. But what if I don’t want the average? If I seek the data point two standard deviations out? If I’m channeling my inner Anthony Bourdain & I’m seeking the fermented shark or cobra heart within my query? At Google, we ranked web pages & ads with many signals. An engineer taught me that of them found rarity of a word across a set of documents. For example, if there’s a document containing 10 instances of the word ghoti across a collection of documents, where the average document has 0 ghotis, that document is likely the best to answer a search about fish.1 I haven’t found a way to goad an LLM to produce the rare result. Why did one user decide to use a particular piece of software when 10 others did not? Maybe I haven’t learned how to prompt an LLM well. Ideogram launched a feature called Magic Prompt that expands a basic prompt into richer instructions that elicit better results, marrying the language of a user & a computer. A user might write : A boy with a dog in a park. Magic Prompt replies : A heartwarming scene of a young boy playing with his happy, wagging dog in a lush, green park. The boy has a playful smile on his face while he tosses a bright yellow ball for his furry friend to fetch. The park is filled with beautiful flowers and trees, creating a serene atmosphere for this bonding moment between boy and dog. Magic Prompt surprised me in three ways. First, unlike Google search where terse queries tend to work very well, LLMs relish verbosity. Less Hemingway & more Faulkner. Second, maybe there’s a prompt for rarity. If you know what it is, please share with the rarity shibboleth. Third, I wonder what other prompt techniques are out there that have yet to be discovered to wrest the best from an LLM. 1 This technique is called TF-IDF. It means Term Frequency Inverse Document Frequency. |
Older messages
The Challenge of the AI Demo
Thursday, September 5, 2024
Tomasz Tunguz Venture Capitalist If you were forwarded this newsletter, and you'd like to receive it in the future, subscribe here. The Challenge of the AI Demo The AI Demo isn't easy. Many
Which Design Era Are We In?
Tuesday, September 3, 2024
Tomasz Tunguz Venture Capitalist If you were forwarded this newsletter, and you'd like to receive it in the future, subscribe here. Which Design Era Are We In? When the internet became popular
The Dislocation Between Public & Private Web3 Markets
Friday, August 30, 2024
Tomasz Tunguz Venture Capitalist If you were forwarded this newsletter, and you'd like to receive it in the future, subscribe here. The Dislocation Between Public & Private Web3 Markets With
Higher Levels of Abstraction
Monday, August 26, 2024
Tomasz Tunguz Venture Capitalist If you were forwarded this newsletter, and you'd like to receive it in the future, subscribe here. Higher Levels of Abstraction Over the weekend, Andrej Karpathy
What Has Your GPU Done For You Today?
Friday, August 23, 2024
Tomasz Tunguz Venture Capitalist If you were forwarded this newsletter, and you'd like to receive it in the future, subscribe here. What Has Your GPU Done For You Today? A year ago, enterprises
You Might Also Like
[VIDEO] Meet the woman who built a $100m+ soda empire
Wednesday, September 18, 2024
Watch her incredible story design-2-header-newsletter Hi there, We're excited to share an incredible new episode of our podcast featuring Allison Ellsworth, the visionary founder of Poppy! 🌟 Meet
🔥 Swipe Right on User Growth!
Wednesday, September 18, 2024
🗓️ You Are Invited! AMA with Sandhya Ganesh, User Growth Product Lead at Tinder
Love To Build? Maybe Don’t Raise Capital.
Wednesday, September 18, 2024
Looking Back On Seedscout's Fundraising Journey ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏
the simple technique that saved my biz
Wednesday, September 18, 2024
Read time: 1 min. 07 sec. Be honest: How distracted do you feel right now? Slack pings, 𝕏 doom scrolling, iPhone buzzing... We live in a world designed to steal our focus. And this sh*t is getting
Ramen Tools and Timesible
Wednesday, September 18, 2024
Comprehensive time and budget tracking tool for teams BetaList BetaList Daily Ramen Tools See what tools other indie makers are using Timesible Comprehensive time and budget tracking tool for teams Too
🌿 What’s Draining Your Productivity? You Might Be Surprised…
Wednesday, September 18, 2024
Staying healthy while working online can be a challenge
There's a new community of SaaS founders growing
Wednesday, September 18, 2024
MicroConf Hey Rob!👋 In just the first few days of its launch, The SaaS Launchpad 🚀 has seen an incredible response from the SaaS community! Over 110 SaaS founders and aspiring entrepreneurs have joined
the future of energy
Wednesday, September 18, 2024
uncover where key players are making their moves, the emerging markets to bet on, and the startups that should be on your radar as we approach 2025 Hi there, Our expert is breaking down emerging energy
Founder Weekly - Issue 654
Wednesday, September 18, 2024
View this email in your browser Founder Weekly Welcome to issue 654 of Founder Weekly. Let's get straight to the links this week. General Non-Obvious Signs of Early Startup Traction — And How to
🏆 He built a $100m ecomm brand
Wednesday, September 18, 2024
And he's teaching you to do the same Hey Friend , You've probably had that dream—the one where you wake up, finally in control of your time, working on your own terms, and building something