Zach Jones

[Zach Jones @zmjones_]

@carlislerainey "don't worry about people stealing your ideas, if they are any good you'll have to cram them down people's throats"

[Zach Jones @zmjones_]

got laid off from meta. if anyone knows any open ds or swe type jobs that would be good for a social scientist please hit me up!

[Yanis Varoufakis @yanisvaroufakis]

When the get-Al-Capone tactic (catching a hyper-criminal on minor charges) is used against a former President, who is also a front-runner in the forthcoming primaries, will re-legitimise Trump's nastiness in the eyes of 50% of Americans and blow up what is left of US democracy.

[Zach Jones @zmjones_]

@Mr080guy @SentientPotato6 ❌ Uni: 47m
✅ Clinic: 13m
❌ School: 20m
❌ Supermarket: 37m
✅ Pharmacy: 13m
❌ Library: 59m
❌ Park: 17m
✅ Bar: 11m
✅ Corner Store: 13m
❌ Gym: 38m
❌ Cinema: 50m

all of these are <20m by bike. 15 by ebike. bikes are the way

[Noah Smith 🐇🇺🇸🇺🇦🇹🇼 @Noahpinion]

Yep. This really nailed it. "Suck on this" was the whole reason we invaded a country that didn't threaten us in any way, resulting in the deaths of 100,000s of innocent people.

[Zach Jones @zmjones_]

@RexDouglass i recall that OSF (or some related group) was working on something like this at one point

[Fridolin Linder @FridolinLinder]

My team at Meta, Core Data Science, is looking for a Ph.D. intern for next summer (London/Tel Aviv).

https://t.co/McEl0pAMuw

Feel free to reach out to me if you are interested or have any questions.

[Zach Jones @zmjones_]

@maya_sen @pkphlam @poliscibitches @profTLe i don't think social science phds are considered luxury goods. a) nobody cares what you got a phd in after you are hired and b) some of roles we are in are measurably impactful and have that going for us

[Zach Jones @zmjones_]

@zeitzoff @Scott_E_Page it did for a while and then died out. it was a hot thing when i took a class with him in like 2015. i think the fundamental problem is we don't have good enough measurement for most really complex phenomena. it is more of a thing in econ i think.

[Michael Weintraub @miweintraub83]

A fantastic perspective, which resonates with me. I have played the guitar, piano, baritone sax, and bassoon, and performed in musical theater. I didn't do any of these well enough to lead me anywhere in particular. But that doesn't really matter!

[Zach Jones @zmjones_]

@Noahpinion the problem with this is adversarial adaptation. most of the policies/enforcement systems can be circumvented and telling users how/why they are being enforced on will just allow them to figure out how to beat the system.

[Zach Jones @zmjones_]

@matt_blackwell i mean i think it does highlight how capricious the whole thing is, which is unsurprising i agree

[Zach Jones @zmjones_]

@rudabagle @CaleyFretz i can't think of a single "classic" ml method that was statistically motivated. most of them still have weak statistical theory. trees, rf, boosting, svm, etc.

[Zach Jones @zmjones_]

@rudabagle @CaleyFretz those are totally different problems. not everything is about causal inference. there are loads of purely detection problems. having good features is important, a/b tests useful for measuring improvement, but they are compliments not substitutes

[Zach Jones @zmjones_]

@rudabagle @CaleyFretz much of modern ml is not motivated by statistical theory. statistical theory comes much later and generally describes a very simplified version of the method

[Zach Jones @zmjones_]

@CaleyFretz i work on the topic what do you want to know?

[Zach Jones @zmjones_]

@rudabagle @CaleyFretz that is nonsense. yes where the data comes from is important as is the framing of the problem. but all of ml/ai is *not* stats. nor is ds just the practice of stats. thought this debate was over in like 2015

[Thomas H. Ptacek @tqbf]

I’m sorry, I simply cannot be cynical about a technology that can accomplish this.

[Zach Jones @zmjones_]

@pudgenet @IDoTheThinking guns exported to mexico from us. guns increasing homicide rate.

[Zach Jones @zmjones_]

@pudgenet @IDoTheThinking also https://t.co/EEL7G7v5dK