Replying toD&D.Sci: The Mad Tyrant's Pet Turtles

I did some initial exploration of the dataset and came to similar conclusions as others on the thread.

I then decided this was a good excuse to finally learn how to use LightGBM, one of the best-in-class tools for creating decision trees, and widely used in the data science industry. In other words, let's make the computer do the fun part!

The goal was to output something like:

If color = blurple: weight is 1234
Else
  If segments > 42: weight is 2345
  Else weight is 3456

What I actually got:

Fangs: ~17 pounds
No fangs: a big tree that outputs in the range of 18-19.5 pounds

I used default settings, transformed color/fangs/nostrils into 0-N categorical variables and

knite4y

Signaling isn't about signaling, it's about Goodhart

TLDR: Honesty is the best policy, and don't be a try-hard.

Replying toSan Francisco shares COVID data only when it's too late

knite4y

San Francisco shares COVID data only when it's too late

I understand that data collection is difficult and empathize with the people responsible for doing the work.

The thing is, SF used to publish everything as soon as they could! We accepted that numbers could be revised up or down as data was fully coded. This 5 day lag is IMO far on the wrong side of timeliness vs correctness.

San Francisco shares COVID data only when it's too late

knite

In the early months of the pandemic, San Francisco posted case and death counts on a daily basis. It was sometimes messy with significant corrections, but the city did the best they could (or at least better than most) to keep the public informed.

Around half a year ago, they got worse. Data is now published with a 5 day lag. Presumably, the politicians and decision makers get timely data. Everyone else has to wait.

This is not okay. This is very not okay. In the post vaccination but pre-Omicron days, I was annoyed philosophically. But now. Now?

https://sf.gov/data/covid-19-cases-and-deaths

Cases as of 12/21, as reported on 12/26

At this point, we're fucked and it's definitely over. But this is still not okay, and now I am furious. Anyone who wants to know the true state of the world cannot, unless they are On The Inside.

So, where is the pebble that starts an avalanche? Who do I need to Tweet or email or call to pressure the city into changing this?

Replying toPaths Forward: Scaling the Sharing of Information and Solutions

knite4y

Paths Forward: Scaling the Sharing of Information and Solutions

Obvious next step: if there's a lot of low hanging fruit like this, let's find it? Have you considered using your LW/Twitter/blog to publicly solicit obvious, simple, and high leverage solutions to other big problems?

Replying toSelf-Integrity and the Drowning Child

knite4y

Self-Integrity and the Drowning Child

In dath ilan, it is virtuous to write more stories about dath ilan.

Replying toObesity Epidemic Explained in 0.9 Subway Cookies

knite5y

Obesity Epidemic Explained in 0.9 Subway Cookies

I expect agave to be generally preferred over table sugar and HFCS due to having a significantly lower glycemic index. I'm unfamiliar with Karo.

Replying toCovid 8/12: The Worst Is Over

knite5y

Covid 8/12: The Worst Is Over

Something I've been wondering for a while: are organizations/journalists/individuals filing FOIA requests to get emails and other relevant documents about how the CDC and FDA made their COVID decisions?

Replying toWho wants to run a B2B SaaS startup?

knite5y

Who wants to run a B2B SaaS startup?

Potentially interested!

Big picture, if your friend wants a different blend of upside-to-work, perhaps they should consider hiring someone to work 15-20 hrs/wk, freeing them up to do <5 hrs/wk of supervision?

Replying toWhy Prefetch Is Broken

knite5y

Why Prefetch Is Broken

This post is a bit hard to parse - please consider replacing "a.test" with something like "test.com/a" or "a.test.com/page" to clarify whether the issue is per-page caching or per-domain caching.

Replying toD&D.Sci April 2021 Evaluation and Ruleset

knite5y

D&D.Sci April 2021 Evaluation and Ruleset

I posted my answer a bit late but this was a ton of fun!

Grading myself on SSC's 2020 predictions

knite

Mantic Monday: Judging April COVID Predictions

Judging Our April 2020 Covid-19 Predictions

I privately recorded my own predictions on these questions. I felt uncomfortable posting these publicly at the time. I can't fix that, but I can post my answers now and do better next time.

Score: 3.34

questions

Edit: Per @bucky, I'm pretty sure the scoring rule I used (ln(p) - ln(.5)) is wrong, but I'm not sure what would be the correct rule.

Best of LessWrong: Jan 10 - Jan 17

knite

I'm an infovore. The failure mode for an infovore is spending too much time reading and not enough time doing. I was bit especially hard by this a few days ago - I hadn't caught up on LW posts in while and had 100+ in my RSS reader. I wanted to read them all but I knew I shouldn't.

I had the same issue with Hacker News for many years. I broke the cycle by signing for for a daily "best of HN" feed, however even this proved to be too much content. I replaced it with a weekly best of digest and that's been working well for me - the digest... (read 282 more words →)

Where are the post-COVID complainers?

knite

Assumption: COVID immunity lasts a long time, because that's how diseases work, and this is the case for 95+% of people who get COVID or a COVID vaccine.

20% of the United States has had COVID (13-30% per https://covid19-projections.com/) and presumed immune. Even if we conservatively take the official case count (20M) and trim off 25% to account for possible duplicates, that's 5% of the population.

Another 1-2% has already been vaccinated and will join them in a month, with more every week after that.

So on the low end, in mid-January, 7% of the country will be fine. COVID-fine, at any rate, but still subject to all local, statewide, and national restrictions.

Why aren't these... (read more)

Ongoing free money at PredictIt

knite

Previously on LW

PredictIt is structurally inefficient

A maximum of 5,000 people in one market
A maximum of $850 in betting per person per contract (not per market)
A substantial house cut of 10% on profits and 5% on withdrawals.

Because of the market's structure, smart money cannot (given the first two reasons) or will not (given the third reason) fix all instances of mis-priced markets.

Free money? Free money!

Have you ever wished you could bet on past events? Well, today is your lucky day!

Q: Who won the Texas Democratic primary? A: Joe Biden

Q: Who won the presidency? A: Joe Biden

As I write this, you can purchase $1 for the price of $.85, in the Will the 2020... (read more)

Current guidance for COVID-19 self-care?

knite

What is the current consensus on supplements to take or skip? I thought zinc lozenges and high dose Vitamin C were clear winners, but I am now finding conflicting information on how Vitamin C may impact interferon.

There's also the HCQ issue, where it's been so politicized that it's rather difficult to determine efficacy as an early intervention.

Edit: The internet exists, I have no plans to defect in the Prisoner's Dilemma and go shopping in person.

Asking for myself, because, well, y'know. Getting tested tomorrow, and regarding that - come on San Francisco, it's nearly September, DO BETTER. Please DM if you know of a way to get tested with a <24 hour turnaround time.

Which COVID-19 serology (antibody) test is best?

knite

https://www.fda.gov/medical-devices/coronavirus-disease-2019-covid-19-emergency-use-authorizations-medical-devices/eua-authorized-serology-test-performance

The FDA lists two dozen tests with different sensitivity and specificity characteristics, different sample sizes, and different 95% CIs. Some of these tests are clear losers (lower bound on sensitivity CI ~75%!), but are there any clear winners?

Maximizing impact at startup with high EA potential?

knite

An early stage startup (SF Bay) has brought me on in a position of technical leadership. The product has the potential to save high-billions of human hours, and eliminate low-billions of tons of carbon emissions. Given that all startups are wildly optimistic, moving over two zeroes still lands at ~X00m hours and ~X0m tons of carbon saved.

It took me three weeks to realize I should run the numbers. After doing that, it took me three minutes to realize I've suddenly found myself in a potentially high impact and high leverage role. And that is precisely why I'm writing this post.

Beyond doing my job to the best of my abilities, what the heck else should I do? Are there EA/LW/rationalist resources and connections I should be contacting? Leadership coaches? Should I be advertising our job openings here? Talking to 80K hours? Playing NDA chicken and sharing more details about what the company is actually doing?

Advice greatly appreciated!

LESSWRONG
LW

LESSWRONG
LW

knite

Current guidance for COVID-19 self-care?

Where are the post-COVID complainers?

Grading myself on SSC's 2020 predictions

Ongoing free money at PredictIt

knite

San Francisco shares COVID data only when it's too late

Grading myself on SSC's 2020 predictions

Best of LessWrong: Jan 10 - Jan 17

Where are the post-COVID complainers?

Ongoing free money at PredictIt

Current guidance for COVID-19 self-care?

Which COVID-19 serology (antibody) test is best?

knite

Current guidance for COVID-19 self-care?

Where are the post-COVID complainers?

Grading myself on SSC's 2020 predictions

Ongoing free money at PredictIt

knite

San Francisco shares COVID data only when it's too late

Grading myself on SSC's 2020 predictions

Best of LessWrong: Jan 10 - Jan 17

Where are the post-COVID complainers?

Ongoing free money at PredictIt

Current guidance for COVID-19 self-care?

Which COVID-19 serology (antibody) test is best?

Previously on LW

PredictIt is structurally inefficient

Free money? Free money!