jon_richards

Does qualitative data actually exist? Named colors are considered qualitative, but rbg and cmyk are quantitative. Does converting from one to the other switch whether it is qualitative or quantitative?

Surely semantic meaning is qualitative, but look at word replacement in Google search. That’s entirely based on statistics, thesaurus graphs, and other ultimately quantitative data.

The neat thing about neural nets is that they are ultimately making a very, very complicated stepwise function. Brains are not neural nets, but are they doing anything other than create a very complex, entirely numerical, time and state dependent function? No matter which way you try to understand something, ultimately you are relying entirely on statistical inference.

I’ve thought about a system where a company has to purchase liability insurance when they dump assets.

So if a plant has a chemical leak, the company can’t sell every other plant, pay a huge dividend, and then declare bankruptcy to avoid paying damages. The plant would have to be insured for up to the total dividend payout. The insurance company would charge a huge fee based on what they think the plant will end up costing after a court battle, etc.

Genuine question: Why is enforcement so difficult? Let everyone declare their primary residence and give property tax credits based on the number of occupants (with pass-throughs for rental properties, etc). Everyone has an incentive to declare a primary residence and it shouldn’t be too difficult to find people declaring multiple.

You could cap the number of occupancy credits per sqft to keep things reasonable.

It would also help with things like state taxes based on primary residence.

Woodrow Wilson’s name got removed from a Princeton college and they stressed it was because he was particularly abhorrent even for his time.

Maybe the same will happen with meat. Veal, foie gras, shark fin, octopus, endangered animals, etc. Most of the blame will fall on producers, but “carnivorous” over eaters like Ron Swanson from parks and rec might age poorly.

Whenever people bring up gay marriage legality, I think it’s also important to show this graph. https://xkcd.com/1431/

Expect live meat to be overwhelmingly abhorred before you see a ban.

This really annoys me in PostgreSQL. Django encourages you to use joins across one-to-one relationships, but the query planner doesn’t know how to apply the limit before the join. It’s relatively simple to change select_related to prefetch_related on all paginated queries, but it still feels like the query planner should do it, especially for left joins (inner joins are slightly more complex to check).

> without a good recommendation from someone I personally know.

I’ve always thought there’s a big opportunity here for principle component analysis or other dimension reduction. The reviews you see could be filtered down to people who have reviewed other things similarly to you, like how Netflix gives different ratings for the same show depending on what else you’ve watched.

It’s almost like establishing trust. If someone has rated that USB hub 5 stars, you probably don’t want their opinion on a restaurant either. It would also give a real incentive to leave honest reviews, which would help collect data from the massive group that currently just doesn’t care.

College classes are often split into lectures and precepts/discussions/office hours. I think the thing holding back your idea is the expectation of being taught in-person by a professor. For large classes, that’s only really possible in a lecture format.

My personal favorite method of learning was a TA that went through the material, meticulously calling on each person in order. You could skip, but it was still obvious if you were playing a game or browsing Reddit. Eventually you learned that you really didn’t have all the answers and that paying attention was probably a good idea. It also got people comfortable giving a wrong answer and that really helped everyone with the learning process.

That’s because you aren’t in free fall.

Subsidies have transitioned away from direct payments to subsidized insurance. The insurance isn’t just about natural disasters or crop loss, it guarantees minimum revenue based on farmed area, basically resulting in direct payments whenever prices are low. That has been successfully rebranded as “not a subsidy”.

https://farm.ewg.org/progdetail.php?fips=00000&progcode=corn

Yes, I was asked by a Junior if they should always use f strings now. There are still a couple situations where you need format, such as when you’re formatting a variable instead of a literal.

Also, shoutout to Template for untrusted strings. It’s surprisingly unknown. Most people suggest replace or a custom format function.

I thought someone was going to raise a flag.

It used to be a common email scam that you’d make two opposite market predictions and send them as free advice. Then you’d do the same thing the next month, but only send to the people who received the “correct” first message. And again. And again. And when you had “proven” yourself all-knowing to a small group of people, you’d start charging them.

Something like 3% of the total land area in California burned this year.

They look a lot like fencing gloves, actually.

Any sufficiently misunderstood algorithm is indistinguishable from AI.

> I was the logical choice. It calculated that I had a 45% chance of survival. Sarah only had an 11% chance. That was somebody's baby. 11% is more than enough. A human being would've known that.

-I, robot

That’s starting to sound like an exoskeleton

I think it’s primarily for people copying and pasting something like an error message, which may have user-specific data in quotes. That should be what “not many results for...” is for, but Google is always trying to optimize away those clicks. Maybe they could add a “programatic” search feature with documented syntax for power users.

If you repeatedly add 2 and then divide by 2, you approach the number 2.

If you repeatedly flip a coin to either add 2 or divide by 2, what distribution do you approach?

If you have a distribution, what random processes are identities for that distribution? Which are stable? Which will reproduce the distribution from any starting distribution?

It’s related to matrix kernels, but it’s hard to generalize to continuous numbers. I spent a while looking into it years ago and couldn’t find much.

My goal was to eventually take it a step farther and create simple stochastic functions for certain behaviors, such as bistable switches, memory registers, etc.

Most of the time, GET should not affect system state at all (other than caching, etc). Idempotent can still have effects on the first access. Even deletion can be idempotent.

I generally see PUT and PATCH classified as idempotent (for the same input).

I think people are picturing chapter 11 bankruptcy, not 7 or 13. The firms stay intact, but the shareholders lose everything and the bond holders come to a new agreement.

Calling once selects for people more likely to fall for the actual scam.

One really important feature of time series data is the preservation of what the dataset looked like at each point in time. Financial data providers will make a mistake (off by order of magnitude, missed a stock split, etc) and then go back and correct it. This means you end up training models entirely on corrected data, but trade based on uncorrected data.

The point is that it’s biologically impossible to test for some things with a finger prick. Theranos was like an electronics company claiming to have “solved” the uncertainty principle. In all these tests people are proposing, you still have to abide by the biological requirements.

The weirdest thing about us is most likely the fact that our moon is almost exactly the same size as our sun (relative to distance).

2.4B is monthly.

I think one issue is modern AI design. Their chance of hitting you goes down with distance and cover. Speed and surprise might be given some consideration, but you are almost always safest peaking around cover and clicking on enemy heads.

I'm not sure if the original Ghost Recon was actually this advanced, but it felt like enemies got more accurate as they got more evidence of your position. You could find a good sniping spot and take out a couple enemies as they fired in your general direction, but the next enemy that came over the ridge (after hearing your shots) would immediately kill you with their first shot. The solution was to use the rest of your squad. Once you got a couple kills on your sniper, you'd hide completely and switch to an assault in another position. The enemy would run over the ridge, but they'd only have improved accuracy against the sniper. That made them an easy fight for the assault.

Translated into the position graphs in the gamesutra link, there was a penalty for fighting from the same node for too long. This results in an inherent incentive to capture new territory and turns the "safe" hallway into a deathtrap.

I’ve always assumed that was also the inspiration for one of the earliest xkcd’s: https://xkcd.com/10/

This is how they sell those magnet bracelets that "align your chakras" or whatever. The reason you can reach farther the second time is that you stretched.