One of my mom’s coworkers, who was a non native English speaker mashed up “People who live in glass houses shouldn’t throw stones,” and “You shouldn’t kick a dead horse,” to “People who live in glass houses shouldn’t kick dead horses.”
One of my mom’s coworkers, who was a non native English speaker mashed up “People who live in glass houses shouldn’t throw stones,” and “You shouldn’t kick a dead horse,” to “People who live in glass houses shouldn’t kick dead horses.”
Head over heels?
Me also!
Court papers filed Wednesday state that Mr. Abrego Garcia was beaten, deprived of sleep, and subjected to psychological torture during the nearly three months he spent in Salvadoran custody. It is unacceptable to send people to other countries to be tortured!
I tend to agree that ranked choice voting is widely preferable to our (general) current system in the US, but I think it's important to note every electoral system has quirks and vulnerabilities and ranked choice doesn't mean bad/incompetent/just plain evil folks can't still win elections/majorities
Chatbots — LLMs — do not know facts and are not designed to be able to accurately answer factual questions. They are designed to find and mimic patterns of words, probabilistically. When they’re “right” it’s because correct things are often written down, so those patterns are frequent. That’s all.
👇
A Friendly Reminder...
NEW: The US government has collected DNA samples from 133,000+ migrant children and teenagers and uploaded them into CODIS, a national criminal database originally built for convicted sex offenders and violent criminals, according to documents reviewed by WIRED.
We really need to remove the pardon power. The president shouldn’t have the power to incite a rebellion against the United States and then pardon the guilty. The same applies to bribes and selling pardons for personal profit.
Wrote a piece for @opensource.org based on my experiences in open source over the last 20 years (!).
A man who holds the title of President of the United States, who took an oath to protect and defend the Constitution, who then says he doesn’t know if he’d uphold the Constitution, is unfit to be president and should be removed.
I agree with the responses saying that it isn’t worth responding, but the comment is just wrong. Even very good programmers make mistakes with C and C++. The first time I used valgrind on my code was quite humbling.
The 5 largest single day $DJIA drops and their corresponding President. There seems to be a pattern here but I can’t quite figure it out.
And yet Joe Rogan puts guests on who say "Vaccines aren't actually responsible for the reduction in infectious diseases.”
Breaking News: The top vaccine official at the FDA abruptly resigned, saying that Robert F. Kennedy Jr.’s “misinformation and lies” were irresponsible.
I think the way to paraphrase would be: “Signal does not put our military personnel at risk. Stupid people not knowing how to use Signal puts our military personnel at risk.”
Right now there are a lot of new eyes on Signal, and not all of them are familiar with secure messaging and its nuances. Which means there’s misinfo flying around that might drive people away from Signal and private communications. 1/
🔑 under-discussed point — why they were on signal:
Corey Doctorow: "Even by Amazon standards, this is extraordinarily sleazy: starting March 28, each Amazon Echo device will cease processing audio on-device and instead upload all the audio it captures to Amazon's cloud for processing, even if you have previously opted out of cloud-based processing"
Happy π Day! It is annoying that I need the Greek keyboard installed on my phone to get the Greek letters. At least the common ones like π should be available as alternatives like ö.
Love Signal. Recently got my mom on it as well.
As I fear not the child with a weapon he cannot lift, I will never fear the mind of a man who does not think. — Wind & Truth This quote seems hopelessly optimistic given how dangerous Trump is. Sigh.
Bluesky really needs to support other reactions than just like and share. I’m🤬 also.
Watching the GOP Congress completely abdicate their constitutional responsibility to control the purse is really something.
An actual scientist-entrepreneur is making waves with his breakup with Musk. I can't vouch for what he writes, but NeuroVigil is a real thing on Round B venture capital funding. www.linkedin.com/posts/dr-phi...
Thank you, President Carter for a life well-lived. You were always an example of the best of us. We’ll take it from here.
To find the Christmas tree for #AdventOfCode day 14, I checked how many robots were in a triangle 🔺 between the lower left corner, the middle of the top, and the bottom right corner. When 75% of the robots are in the triangle, you have the Christmas tree. github.com/omalley/adve...
That means I don't need to redo the previous stuff, because I start right at the new block. I then pop off the path back to the block and do the previous location.
It all should be translatable to C++. I basically created a small (max 4 items) stack for each place in the grid. When I move to a new square, I push my previous position to new location's stack. That lets me unwind my path and detect loops. The code iterates backwards and puts a block on the spot.
Day 12 of #AdventOfCode was fun because I got to revisit the union-find algorithm that I learned in grad school. (I found a #RustLang crate that implemented it for me, which was great.) Counting the edges of the shapes was challenging. My solution runs in 0.734 millis. github.com/omalley/adve...
I've been having fun trying to optimize day 6 on #AdventOfCode, which is my slowest day. My original implementation was 590 millis. Changing the data structure to minimize repeated work took it to 55 millis. Moving from Vec> to Array2 saved another 13%. github.com/omalley/adve...
In #AdventOfCode day 11 part 2, the problem description doesn't give the output on the test data. I *hate* when they don't give a test for part 2. Anyways, the expected output for part2 on "125 17" is 65601038650482. github.com/omalley/adve...
I really enjoyed watch this talk about #AdventOfCode. I used Advent of Code to learn Rust a few years ago. It has been great fun doing the puzzles and comparing notes with my friends.
Part 2 of day 5 of #AdventOfCode took a while to get correct. It is an interesting tradeoff in how much to setup your data structures. (It reminds me of studying in university and how much to memorize versus re-derive.) In this case, I sorted and grouped the rules. github.com/omalley/adve...
Nice catch. You're right. I've fixed it now and added a test case. github.com/omalley/adve...
If you want to compare with my solution, which also doesn't use regex, look here - github.com/omalley/adve... . The key observation is that Peekable just gives you a one character look ahead and you need decide what to do at each character point. Look at the handling of parsing "do()" and "don't".
My regex version is in the git history. You can see it here: github.com/omalley/adve...
I did it both ways. The regex approach was a lot less code (36 versus 103 loc), but the manual parser ran much faster. (2.3 ms versus 0.078 ms). github.com/omalley/adve...
Yikes. It is far to easy to recompile the regexs over and over again. Fortunately, #RustLang 's linter, clippy, has a check for compiling regexs in a loop. Of course, it doesn't catch all of the cases.
I've thought about using #RustLang 's macros to compile the regex to a state machine at compile time. That would remove the startup time that makes it hard to use in very fast applications like #AdventOfCode where I'm trying to get the run time for all 25 days to less than a second.
If you are trying to just solve the problem, regexs are great. The advantages of the manual parser are mostly no start up time and the fact that you have functions to create abstractions. You do end up with more code to validate and maintain.
Day 3 of #AdventOfCode, I first used regexs with 31 lines of code for a total run time of 2.26ms. Moving to manually parsing with a peekable char iterator with 103 lines of code took 78µs. So 3x the code for 29x speed up. (Compiling the regexs at runtime is slow.) github.com/omalley/adve...
In day2 of #AdventOfCode, my first pass of part 2 was relatively complicated. Simplifying it down to just manually trying to delete each item in the row was far simpler and faster. github.com/omalley/adve... Day 2 (160.29µs) · Generator (111.83µs) · Part 1 (8.92µs) · Part 2 (39.54µs)
If you want to see my code for the days so far, it is at github.com/omalley/adve... . I try to leave the git history accurate if you want to see what kinds of clean ups I do over time.
I really enjoy doing each year's #AdventOfCode, which is a set of daily programming puzzles each December. Although their is a leader board for getting the answers the fastest, I prefer to write nice understandable and fast code. Using #RustLang I can usually get all 25 problems running in under 1s.