The Ledger
Every statement that passed quotation and attribution checks. Mix any filter with any other: certainty 1/5, debate potential 5/5, or both at once.
why aren't all 2,374 resolved? a statement only gets an assessment when the public record can support or contradict it. opinions and what-ifs never can, and 6 checkable ones are still open, waiting for their date. predictions held up or didn't; assertions are supported or contradicted. on every card: ▮▮▮▮▮ certainty · ▮▮▮▮▮ debate potential. speakers are clickable
a16z data shows top revenue startups almost always overlay direct sales
“If we go through our portfolio, if we go through the data that we have, we look at the companies that actually hit the most revenue, and in almost every case, they overlaid sales on top of the organic growth.”
Mayo Clinic doctors lack clinical data as rich as K Health's
“And just to be very clear, a doctor in Mayo Clinic does not have access to this information. There's no medical grade system that allows you to compare yourself to a data set that's as rich and relevant as you are.”
DeepMind's reinforcement learning fails on minor visual changes while Vicarious succeeds
“So if you use, ah, if you compare, ah, Vicarious, ah, system with, ah, DeepMind system, ah, for example, if you change the brightness of the screen DeepMind system will stop playing, ah, because it will, ah, get confused, ah, and if you offset the paddle we ou…”
Early Apache Spark was so unstable that clusters crashed within two hours
“At the time we made that bet, Spark was barely working. I mean, seriously. It was a really promising technology, very complete programming model. We loved the fact it was in memory, but we couldn't find a single customer that could keep their cluster running f…”
Kulkarni: 3 TimescaleDB nodes outperformed 30 Cassandra nodes at 1/10th cost
“For time series data, three time scale nodes was able to outperform a 30 node Cassandra cluster at one 10th the cost.”
Spotify listening data can accurately predict user personality traits
“We actually had some sort of collaborations with academic institutions, and, you know, we can actually pretty accurately predict people's personality just based on the data, right?”
Two AI-generated ALS compounds outperformed standard-of-care treatments in lab validation
“We actually put the five hypothesis in the lab, and at this point, it's in vitro and in vivo validation, and, you know, two of them didn't work well one of them work as well as this kind of standard of care at the moment, and two of them work actually much bet…”
Top machine learning talent is heavily concentrated at Google and Facebook
“Almost all of the really smart machine learning people are in less than 10 companies in the world. They're actually in less than five, right? In fact, they're mostly all at Google and Facebook.”
IBM Watson's $60M MD Anderson oncology project was a complete failure
“Now several years later, and sixty million dollars down the drain, that initiative failed. It actually failed.”
Twitter and LinkedIn's open-source data catalogs go unused without governance
“Reason why I say that is because I've talked to all the high-tech companies like Twitter, LinkedIn, et cetera, and they all have their catalog projects. Many of them have actually open sourced catalog initiatives, But then you find that nobody's actually using…”
Altshuler: Mathematical invariances guaranteed in all structured human data
“There is a set of mathematical invariances that Are guaranteed to happen in any structured human data.”
Golde: Via delivers two passengers per mile, outperforming competitors 3-4x
“So, on every mile we drive, we actually deliver on average two people one mile towards their destination, which is like four times better than You know, four times better than taxis, more than one in a, ah, more, more than three times better than, kind of, the…”
David Loaiza: Roughly 80% of alternative data sold by vendors is duplicated
“You see people taking data from that set, and you compare it, it's about 80% the same data that someone is selling it.”
Uber dropped Google Maps for Foursquare to power venue search
“If you use Uber and you say, like, I, you know, please pick me up at the, you know, Regency Cinema in Tribeca and take me to the Starbucks on 27th street, rather than typing addresses, That's all powered by Foursquare's global data set and they moved off Googl…”
Foursquare data showed women's foot traffic to Trump properties dropped 29%
“During the peak of the campaign, Women, women's foot traffic in blue states to Trump branded properties was down 29% year over year.”
Glueck: Bernie Sanders' campaign used Foursquare data to target Michigan voters
“The Sanders campaign used the data in Michigan, which was the primary that they won, to sort of figure out likely voters, and so we know who goes to colleges. We know, you know, which phones go to yoga studios and buy Subarus, you know, and so, like, we were a…”
Wickramasekara: Clunky enterprise systems force pharma scientists to email sensitive IP
“A lot of these companies, they do have systems that exist, but they're quite clunky and not integrated. So the scientists resort to using email, and this is whatever email systems they have set up. Sometimes they're personal gmails as well.”
Bordes: 2016 image recognition AI uses Yann LeCun's 1993 neural architecture
“This is exactly the same architecture of what's being used right now.”
Dauber: Portfolio AI analyzes medical cases 10% more accurately than humans
“We have a portfolio company in the med tech space where, you know, my dad was a pathologist, and when I was a kid, it would take him 20 minutes to analyze a case to determine whether or not you had cancer. You know, I think in a minute, ah, these guys can anal…”
Chung: Tech giants offer seven-figure salaries to undergraduate AI recruits
“If you go straight to academia and try to pull them out of, even if it's, you know, undergrad, they're getting seven figure offers from Google, Facebook to go work on, whether it's core, Deep learning tech, AI, or better ad targeting.”
Job listings with over 50% bullet points significantly reduce female applicants
“If you go above 50% bulleted content in a listing, you quickly reduce the proportion of women who are likely to apply for the job, statistically.”
MongoDB's AGPL license prevented AWS from offering it as a managed service
“It's an open source license, it's a free software foundation approved license, but it's the AGPL, and the AGPL gives us a little bit more advantage than other players, so Amazon can't quite do the same thing that they can with, ah, MySQL as they can with, so t…”
Amazon Web Services holds about 90% market share
“You have a company that's gonna, ah, it's on its path to be, and it is, about a 90% market share player.”
Mehta: The top B2B customer retention predictor is CEO friendship
“So in our business kind of predicting churn, you know what the number one predictor of customers retaining is? If the customer's CEO is friends with the vendor's CEO, right?”
Falkowitz: Area 1 Security has captured the entire internet for security
“What we've been able to do is capture the entire Internet, right, in a way that maybe you would think Google or others do, right, and to do that for a very specific sort of security purpose.”
Area 1 Data: Top spam filters miss 9 targeted phishing emails per million
“Our data shows that for every one million messages a company receives, nine of them are of a targeted and sophisticated nature that the best spam, Google spam, others are going to miss, right?”
JPMorgan is avoiding blockchain and digital currency, unlike Goldman Sachs
“When you compare that, for instance, to JP Morgan, you don't, you see JP Morgan has sort of avoided the digital currency space, right? They're not doing anything in blockchain or Bitcoin you know, like Goldman is, right?”
Muglia: Dynamic compute scaling is impossible with Amazon Redshift
“In fact, it's possible to change the size dynamically on the fly based on the kinds of queries you're running. And none of that is possible with a traditional data warehouse running on premises or even a traditional data warehouse like Redshift running within,…”
Muglia: Many early Snowflake customers converted directly off Amazon Redshift
“Virtually all of our customers have looked at Redshift. And quite a few of our customers are converted off of it, because we're very fundamentally different than they are.”
Howie Liu: 90% of spreadsheets are used as makeshift databases
“90% of spreadsheets are not used for kind of their intended purpose of numerical analysis, but really, you know, they're used as a makeshift database.”
Kaganovich: Genetics companies build clinical DNA tests using unmanaged public FTPs
“Every genetics company, every hospital, they design and implement tests To figure out what disease you have, or don't have, or whether you should have a child, or what drug to take based on your DNA. They use this data that's collected by academics and deposit…”
Bisignano: Science has not truly sequenced the entire human genome
“We actually haven't truly sequenced the whole human genome just because there's so much variation that we can't capture yet.”
Gutman: 74% of nighttime ER visits are unnecessary
“74% of people that go to ERs at night don't need to be there.”
LeCun: Backprop designs vision systems better than human engineers
“In fact, those dumb learning algorithms, Backprop, namely, is better at designing vision systems than very experienced engineers and scientists are at designing vision systems.”
LeCun: Deep learning designs NLP systems better than linguists
“You know, those deep learning algorithms are better at designing NLP systems than linguists are, or natural language specialists.”
Financial traders unknowingly rely on manipulated maritime tracking data
“So today, traders, quant funds, use data that is being increasingly manipulated, increasingly unreliable, and they have no idea this is happening.”
Automated Insights generated more content than all media combined in 2013
“We generated more content than all the media companies combined last year. Again, three hundred million.”
Unmodified on-premise Hadoop distributions fail in the cloud beyond 10 nodes
“If you just take a normal Hadoop distro and try to run it in the cloud, the chances are at 10 nodes it'll work fine as you start growing and, you know, as you start growing and growing and growing further. Things will start breaking because, you know, compute …”
Gourley: Algorithms can briefly wipe $1 trillion from financial markets
“Sometimes the algorithms get together and decide collectively to wipe about a trillion dollars of market capitalization off the global financial markets only to return it about 15 minutes later with nothing really happening in the world around that.”
Sisense's system gets faster as more people use it concurrently
“So the outcome is actually that the more you use the system, the faster it gets. So this is a system that gets faster with more users, not slower.”
Bloomberg and LexisNexis failed at structuring legal data
“Bloomberg tried to do a lot of the things that we were doing and did not succeed. Lexus has tried to do a lot of these things and has not succeeded. It's not something where you just throw resources at it.”
eHarmony couples had nearly half the divorce rate of offline couples
“The couples that met on eHarmony also had a low divorce rate, so nearly half of what the marriages that met actually offline.”
Smolan: Humanity is at the cusp of re-engineering the human species
“I mean, we're at the cusp of being able to re-engineer our species. This is the first time in human history we've ever been able to do this.”
Mike Driscoll: Telcos hold the world's most powerfully predictive social graphs
“If you want to know who has the most powerfully predictive social graphs on the planet, it's the telcos, because who you call and who you text is an extraordinarily strong indicator of who you actually are connected to in a social way.”
Socher: Grounding models in search results largely resolved factual hallucination
“And so I think the initially people thought, oh, we need neuro symbolic reasoning, blah, blah, to do all this. We just needed more examples of don't hallucinate now, like, take real facts from a search engine, and then mostly summarize those. And then those, l…”
Socher: Profluent closed multi-billion dollar contracts with Eli Lilly
“They've now closed, like, multi-billion dollar contracts with Eli Lilly at Profil and his company because they've created new kinds of proteins that are, for instance, even better than CRISPR-Cas-Nine and at gene editing and being even more specific and target…”
Socher: Parallel Bio secured FDA approval to skip certain animal trials
“That company alone has already gotten FDA approval to skip certain animal trials.”
Wolf: Frontier AI Training Has Shifted From RLHF to Pure RL
“What we know though, is we moved from this pure, like human data, you know, that was first just pre-training on human data and then also aligning with like human preferences that was called RLHF, where we had a lot of human in the loop and human data. To like …”