Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
jwilber
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
17 ms
·
181.
▲
by
jwilber
4y ago
I don’t think they made it sound like that, they just used the word, if you don’t know the definition it’s on you?
182.
▲
Beenz: Bitcoin Predecessor
(en.wikipedia.org)
2 points
by
jwilber
4y ago
|
0 comments
183.
▲
ChatGPT wrote an attack plan for an incel, mentioned specific schools
(twitter.com)
1 points
by
jwilber
4y ago
|
0 comments
184.
▲
by
jwilber
4y ago
You can fit a state space model to missing data in many api’s, sure, but that doesn’t mean it’s supported in any meaningful way. Imagine you have a new product with 6 months of missing data. You can feed that into ets/arima/what
185.
▲
by
jwilber
4y ago
Seems like these guys just wasted $11k to erroneously claim, “deep learning bad! Simple is better!” There’s definitely use for these classical, model-based methods, for sure. But a contrived comparison claiming they’re king is just misinfor
186.
▲
by
jwilber
4y ago
Or, and hear me out here for a second, people go where they’re most valued. Why would someone volunteer to go work somewhere where they’ll get less money, have less flexibility, be treated as second-class employees, have less work perks, et
187.
▲
by
jwilber
4y ago
Wouldn’t call that obvious - a whole host of reasons exists as to why one wouldn’t rejoin a company that just fired them, many not related to money.
188.
▲
by
jwilber
4y ago
Here’s just what Twitter has open-sourced: https://github.com/orgs/twitter/repositories Laughable that you think 10 people can rewrite all of this in a month. It would take longer to scope out the design docs alon
189.
▲
by
jwilber
4y ago
What about twitter engineers makes you think this?
190.
▲
by
jwilber
4y ago
Hard to take your views earnestly when you write: “The officer in the training exercise that killed Officer Tipping has the same name as the officer in Tipping’s report, but they are not sure if it is the same person.” Talk about grasping f
191.
▲
by
jwilber
4y ago
I guess because what I wrote seemingly doesn’t agree with your viewpoint I am all of those things. Not sure what the cell phone bit is about.
192.
▲
by
jwilber
4y ago
Playing a Smurf vs playing someone at your skill level in competitive is a completely different experience. An enemy widow sniping your entire team over-and-over again is not the same as going overtime and nearly drawing a payload nap. I’m
193.
▲
by
jwilber
4y ago
They only except certain providers, so no Google phone, prepaid, etc.
194.
▲
by
jwilber
4y ago
Makes sense. Valve already does this for all of their competitive games, it’s basically what you need to do to stop easy smurfing if your game is free-to-play.
195.
▲
by
jwilber
4y ago
Somehow I think being compared to PG is exactly the sort HN-specific thing Tom wouldn’t be too keen on. (Not saying it’s bad, just made me chuckle).
196.
▲
by
jwilber
4y ago
“An analogy that flips the script here is that some machine learning techniques, like “deep” neural networks, are perhaps as opaque as human thinking. They do amazing things. But no one yet understands exactly how.” Not to derail the thread
197.
▲
by
jwilber
4y ago
I’ve never used mastodon, but it’s hard to take your judgements at any value when it seems like you haven’t either.
198.
▲
by
jwilber
4y ago
You can save the song and dance. You’re not convincing any home-owners (or even potential renters) with your savior-complex lording. Just be honest. You bought an asset you expect to increase in value, and renting allows you an easy manner
199.
▲
by
jwilber
4y ago
Outlaw Ocean is an absolutely amazing book with chapters on different crazy occurrences in our legally-ambiguous oceans (slavery, mobile abortion ships, mercenaries, repo-men, activists , etc.) Definitely worth a read if this interests you:
200.
▲
by
jwilber
4y ago
That would depend on the veracity of the negative reviews, many of which offer no substance apart from, “it’s too woke and not canon!” I don’t think the show is incredible, but visually the show is pretty awesome, and I enjoy the story eno
201.
▲
by
jwilber
4y ago
The worst thing on GitHub are those “rating” badges in profiles that always say A+ regardless of the profile quality. Just having one of those is a decent screening filter itself.
202.
▲
by
jwilber
4y ago
It sounds like you don’t purchase cars much if you think the price difference is only a few hundred dollars. I went to look for a car recently and, ignoring the Covid shortage fee increase, they added a mandatory $1200 for some bs paint pro
203.
▲
by
jwilber
4y ago
Cruelty is worldwide. See: Rwanda, Yemen, former Yugoslavia, Mexican cartels, ISIS, etc. for a small sample of non-US cruelty in the twenty-first century.
204.
▲
by
jwilber
4y ago
Maybe the current bleak situation of expensive education, unaffordable housing, medical debt, etc. is pushing many in gen z to go for what they see (whether it’s true or not) as their only option?
205.
▲
by
jwilber
4y ago
Your argument reads as, “data scientists only know machine learning! Not me, I use popular models from statistics too!” Not trying to sound offensive, but it’s the same aggressive stance repeated on twitter by every other stats/econ ph
206.
▲
by
jwilber
4y ago
Calling StarCraft II “just deathballs” is a complete misrepresentation and utterly false. Maybe Protoss in HotS/WoL (ie 5-11 years ago) but not anything now. Also drones are the Zerg’s worker unit, the things that fly out of carriers a
207.
▲
by
jwilber
4y ago
This is a great mention, ty!
208.
▲
by
jwilber
4y ago
If you’re interested in how tree-based models work, I wrote an interactive explanation on decision trees here: https://mlu-explain.github.io/decision-tree/ and random forests here: https://mlu-explain.github
209.
▲
by
jwilber
4y ago
The interviews are definitely hard, but I can confirm the recruiters will reach out to you directly if they find your profile interesting.
210.
▲
by
jwilber
4y ago
I like the website! There is no “testing” going on here, unfortunately. You’re just taking turns placing users in the treatment or control group, arbitrarily, depending on when they visit the page. How can you measure any treatment effect w
More ›