Mathnerd314

10 ms·

91.
by Mathnerd314 2y ago
https://fahrplan.events.ccc.de/congress/2024/fahrplan/talk/H... looks fun. Also https://fahrplan.events.ccc.de/congress/2024/fahrplan/talk/P...
92.
by Mathnerd314 2y ago
> we find that poverty is still 3 percentage points higher in treated counties 10 years after the Walmart opening. So... 10 years after Walmart opens, poverty is 3% higher, and annual household earnings decline by $4,230. Is this a huge
93.
by Mathnerd314 2y ago
I almost forgot about that ARM-Qualcomm dispute, the fireworks are only a few days away.
94.
by Mathnerd314 2y ago
Full opinion: https://www.courtlistener.com/opinion/10289420/an-opinion-wa...
95.
by Mathnerd314 2y ago
Well, look at the process of training a chatbot: - first you make a "raw" corpus, with all the information needed to produce an answer - then you generate sample question-answer pairs - then you use AI to make better questions and
96.
by Mathnerd314 2y ago
There is a longer essay called "On Truth", also by Frankfurt, less read but also more interesting IMO.
97.
by Mathnerd314 2y ago
One way to look at it is that most college textbooks are bad. They are written by one person, who has limited time, and limited understanding. And then you have a teacher who knows even less than the textbook, trying to explain the textbook
98.
by Mathnerd314 2y ago
I take away a different lesson: Dart is most likely slow out of the box. The author lists several reasons: int vs. Int64, GC interrupts, load hoist optimization. All of these are issues with Dart. Hence microbenchmarks, even without interpr
99.
by Mathnerd314 2y ago
The actual author is one person, user '5225225'
100.
by Mathnerd314 2y ago
There is the TARDIS monad in Haskell https://hackage.haskell.org/package/tardis-0.5.0/docs/Contro... It doesn't have the multiple timelines or killing features - it just deadlocks if there is a paradox o
101.
by Mathnerd314 2y ago
Yeah, personally I'd use markdown too, at this point it is easier to use llama-3.1-8b to parse markdown / text into your JSON format of choice than it is to massage recipes into a specific markup like Cooklang.
102.
by Mathnerd314 2y ago
These authors have big Google Docs of evidence, https://jonathanhaidt.com/reviews/ . But if you read it, you will see the effect is (AFAICT) limited to certain populations. There is a significant fraction of students th
103.
by Mathnerd314 2y ago
> There’s a proper way to do this. Is there? Seems like people will complain however fast you roll out AI, so you might as well roll it out quickly and get it over with.
104.
by Mathnerd314 2y ago
I looked into doing something like this and I was like "Todoist has AI, Notion has AI, how am I going to do any better than them?" So I didn't start. At this point I have specific prompts for ChatGPT and when I need this func
105.
by Mathnerd314 2y ago
Mozilla has a list, https://publicsuffix.org/list/ , relatively easy to update. I'm sure there is some Python wrapper library they could use.
106.
by Mathnerd314 2y ago
I would say to look at credit utilization as a function of the other variables. Or more generally each variable as a function of the other variables. There is a bit of this in the initial correlation analysis, but a GLM is more robust.
107.
by Mathnerd314 2y ago
I think glm would be more interesting than clustering, would be good to compare the two. Or maybe UMAP.
108.
by Mathnerd314 2y ago
> Always Review AI-Suggested Code So really the problem is a lack of code review... but I seem to recall that AI is decent at code review too. It won't spot state machine bugs, but SQL injections, no problem.
109.
by Mathnerd314 2y ago
This is one of the common accidents, the "right hook". https://velosurance.com/blog/how-to-avoid-most-common-riding... https://www.fcgov.com/traffic/pdf/bicycle-crashes.pdf I mostly
110.
by Mathnerd314 2y ago
> Free variables and closures Are these even types? I always mentally filed closures under "implementation detail of nested functions".
111.
by Mathnerd314 2y ago
It doesn't account for quantum computing? Cracking passwords seems like one of those things that should get an exponential speedup with quantum computing.
112.
by Mathnerd314 2y ago
But Google didn't pay 2 billion to the couple - it went to the EU government. So now they have a civil lawsuit...
113.
by Mathnerd314 2y ago
> Living in a non-narrative way means rejecting a particular identity, and instead seeing life and meaning as a set of open choices. I choose to fly up through the sky into outer space, away from this place. ;-) Oh wait, I don't hav
114.
by Mathnerd314 2y ago
For reference, SuperTux's code: https://github.com/SuperTux/supertux/blob/8d79ac7ce4db5e4225... (disclosure: not code I wrote, but I've tweaked it here and there) The idea is similarly to just simul
115.
by Mathnerd314 2y ago
This seems more like Fermi estimation ( https://en.wikipedia.org/wiki/Fermi_problem ). And it is underrated and also it is really hard. I mean, it is easy for the author here to say "look I threw it into Google and
116.
by Mathnerd314 2y ago
Actual study: https://www.nature.com/articles/s43587-024-00702-3#Sec2
117.
by Mathnerd314 2y ago
I guess it depends on the study. If it is just comparing between groups, the conclusions probably still hold if they consistently measured blood pressure in the "incorrect" way. If it is something like "85% of Americans have
118.
by Mathnerd314 2y ago
You can read the discussion yourself: https://en.wikipedia.org/wiki/Wikipedia:Articles_for_deletio... The result was "no consensus". That in was 2006, before current policies were well-established, and at thi
119.
by Mathnerd314 2y ago
"Bailout: How Washington Abandoned Main Street While Rescuing Wall Street", Neil Barofsky, page 157 Note though that he is only quoting "off the record" conversations with Treasury Secretary Tim Geithner
120.
by Mathnerd314 2y ago
Seems like basic "you expected that, didn't you?" sort of findings, although they did verify the absence of a lot of correlations. But it's kind of cool that they can directly measure how much impairment a bad night&#x27

More ›